Fourier ptychographic imaging systems, devices, and methods

Abstract

Systems, devices, and methods of Fourier ptychographic imaging configured for illuminating a specimen being imaged from a plurality of incidence angles, for acquiring variably-illuminated, low-resolution intensity images of the specimen, and for reconstructing a high-resolution image of the specimen by iteratively determining the high resolution image that is self-consistent with the variably-illuminated, low-resolution intensity images, for example, by updating overlapping regions of variably-illuminated, low-resolution intensity images in Fourier space.

Description

BACKGROUND OF THE INVENTION

Embodiments of the present disclosure generally relate to wide field-of-view, high-resolution digital imaging techniques. More specifically, certain embodiments relate to Fourier ptychographic imaging (FPI) devices, systems and methods for wide-field, high-resolution imaging.

The throughput of a conventional imaging platform (e.g., microscope) is generally limited by the space-bandwidth product defined by its optical system. The space-bandwidth product refers to the number of degrees of freedom (e.g., number of resolvable pixels) that the optical system can extract from an optical signal, as discussed in Lohmann, A. W., Dorsch, R. G., Mendlovic, D., Zalevsky, Z. & Ferreira, C., “Space-bandwidth product of optical signals and systems,” J. Opt. Soc. Am. A 13, pages 470-473 (1996), which is hereby incorporated by reference in its entirety. A conventional microscope typically operates with a space-bandwidth product on the order of 10 megapixels, regardless of the magnification factor or numerical aperture (NA) of its objective lens. For example, a conventional microscope with a ×20, 0.40 NA objective lens has a resolution of 0.8 mm and a field-of-view of 1.1 mm in diameter, which corresponds to a space-bandwidth product of about 7 megapixels. Prior attempts to increase space-bandwidth product of conventional microscopes have been confounded by the scale-dependent geometric aberrations of their objective lenses, which results in a compromise between image resolution and field-of-view. Increasing the space-bandwidth product of conventional imaging platforms may be limited by: 1) scale-dependent geometric aberrations of its optical system, 2) constraints of the fixed mechanical length of the relay optics and the fixed objective parfocal length, and/or 3) availability of gigapixel digital recording devices.

Some attempts to increase the spatial-bandwidth product using interferometric synthetic aperture techniques are described in Di, J. et al., “High resolution digital holographic microscopy with a wide field of view based on a synthetic aperture technique and use of linear CCD scanning,” Appl. Opt. 47, pp. 5654-5659 (2008); Hillman, T. R., Gutzler, T., Alexandrov, S. A., and Sampson, D. D., “High-resolution, wide-field object reconstruction with synthetic aperture Fourier holographic optical microscopy,” Opt. Express 17, pp. 7873-7892 (2009); Granero, L., Micó, V., Zalevsky, Z., and García, J., “Synthetic aperture superresolved microscopy in digital lensless Fourier holography by time and angular multiplexing of the object information,” Appl. Opt. 49, pp. 845-857 (2010); Kim, M. et al., “High-speed synthetic aperture microscopy for live cell imaging,” Opt. Lett. 36, pp. 148-150 (2011); Turpin, T., Gesell, L., Lapides, J., and Price, C., “Theory of the synthetic aperture microscope,” pp. 230-240; Schwarz, C. J., Kuznetsova, Y., and Brueck, S., “Imaging interferometric microscopy,” Optics letters 28, pp. 1424-1426 (2003); Feng, P., Wen, X., and Lu, R., “Long-working-distance synthetic aperture Fresnel off-axis digital holography,” Optics Express 17, pp. 5473-5480 (2009); Mico, V., Zalevsky, Z., García-Martínez, P., and García, J., “Synthetic aperture superresolution with multiple off-axis holograms,” JOSA A 23, pp. 3162-3170 (2006); Yuan, C., Zhai, H., and Liu, H., “Angular multiplexing in pulsed digital holography for aperture synthesis,” Optics Letters 33, pp. 2356-2358 (2008); Mico, V., Zalevsky, Z., and García, J., “Synthetic aperture microscopy using off-axis illumination and polarization coding,” Optics Communications, pp. 276, 209-217 (2007); Alexandrov, S., and Sampson, D., “Spatial information transmission beyond a system's diffraction limit using optical spectral encoding of the spatial frequency,” Journal of Optics A: Pure and Applied Optics 10, 025304 (2008); Tippie, A. E., Kumar, A., and Fienup, J. R., “High-resolution synthetic-aperture digital holography with digital phase and pupil correction,” Opt. Express 19, pp. 12027-12038 (2011); Gutzler, T., Hillman, T. R., Alexandrov, S. A., and Sampson, D. D., “Coherent aperture-synthesis, wide-field, high-resolution holographic microscopy of biological tissue,” Opt. Lett. 35, pp. 1136-1138 (2010); and Alexandrov, S. A., Hillman, T. R., Gutzler, T., and Sampson, D. D., “Synthetic aperture Fourier holographic optical microscopy,” Phil. Trans. R. Soc. Lond. A 339, pp. 521-553 (1992), which are hereby incorporated by reference in their entirety. Most of these attempts use setups that record both intensity and phase information using interferometric holography approaches, such as off-line holography and phase-shifting holography. The recorded data is then synthesized in the Fourier domain in a deterministic manner.

These previous attempts to increase spatial-bandwidth product using interferometric synthetic aperture techniques have limitations. For example, interferometric holography recordings typically used in these techniques require highly-coherent light sources. As such, the reconstructed images tend to suffer from various coherent noise sources, such as speckle noise, fixed pattern noise (induced by diffraction from dust particles and other optical imperfections in the beam path), and multiple interferences between different optical interfaces. The image quality is, therefore, not comparable to that of a conventional microscope. On the other hand, the use of an off-axis holography approach sacrifices useful spatial-bandwidth product (i.e., the total pixel number) of the image sensor, as can be found in Schnars, U. and Jüptner, W. P. O., “Digital recording and numerical reconstruction of holograms,” Measurement Science and Technology, 13, R85 (2002), which is hereby incorporated by reference in its entirety. Another limitation is that interferometric imaging may be subjected to uncontrollable phase fluctuations between different measurements. Hence, a priori and accurate knowledge of the specimen location may be needed for setting a reference point in the image recovery process (also known as phase referring). Another limitation is that previously reported attempts require mechanical scanning, either for rotating the sample or for changing the illumination angle. Therefore, precise optical alignments, mechanical control at the sub-micron level, and associated maintenances are needed for these systems. In terms of the spatial-bandwidth product, these systems present no advantage as compared to a conventional microscope with sample scanning and image stitching. Another limitation is that previous interferometric synthetic aperture techniques are difficult to incorporate into most existing microscope platforms without substantial modifications. Furthermore, color imaging capability has not been demonstrated on these platforms. Color imaging capability has proven pivotal in pathology and histology applications.

In microscopy, a large spatial-bandwidth product is highly desirable for biomedical applications such as digital pathology, haematology, phytotomy, immunohistochemistry, and neuroanatomy. A strong need in biomedicine and neuroscience to digitally image large numbers of histology slides for analysis has prompted the development of sophisticated mechanical scanning microscope systems and lensless microscopy set-ups. Typically, these systems increase their spatial-bandwidth product using complex mechanical means that have high precision and accurate components to control actuation, optical alignment and motion tracking. These complex components can be expensive to fabricate and difficult to use.

Previous lensless microscopy methods such as digital in-line holography and contact-imaging microscopy also present certain drawbacks. For example, conventional digital in-line holography does not work well for contiguous samples and contact-imaging microscopy requires a sample to be in close proximity to the sensor. Examples of digital in-line holography devices can be found in Denis, L., Lorenz, D., Thiebaut, E., Fournier, C. and Trede, D., “Inline hologram reconstruction with sparsity constraints,” Opt. Lett. 34, pp. 3475-3477 (2009); Xu, W., Jericho, M., Meinertzhagen, I., and Kreuzer, H., “Digital in-line holography for biological applications,” Proc. Natl Acad. Sci. USA 98, pp. 11301-11305 (2001); and Greenbaum, A. et al., “Increased space-bandwidth product in pixel super-resolved lensfree on-chip microscopy,” Sci. Rep. 3, page 1717 (2013), which are hereby incorporated by reference in their entirety. Examples of contact-imaging microscopy can be found in Zheng, G., Lee, S. A., Antebi, Y., Elowitz, M. B. and Yang, C., “The ePetri dish, an on-chip cell imaging platform based on subpixel perspective sweeping microscopy (SPSM),” Proc. Natl Acad. Sci. USA 108, pp. 16889-16894 (2011); and Zheng, G., Lee, S. A., Yang, S. & Yang, C., “Sub-pixel resolving optofluidic microscope for on-chip cell imaging,” Lab Chip 10, pages 3125-3129 (2010), which are hereby incorporated by reference in their entirety.

BRIEF SUMMARY OF THE INVENTION

Embodiments of the present disclosure provide Fourier ptychographic imaging (FPI) methods, devices, and systems for wide-field, high-resolution imaging as used in applications such as, for example, digital pathology, haematology, semiconductor wafer inspection, and X-ray and electron imaging. An example of an FPI device is a Fourier ptychographic microscope (FPM), which may also be referred to as employing non-interferometric aperture-synthesizing microscopy (NAM).

In some embodiments, an FPI system includes a variable illuminator, optical element, radiation detector, and a processor. The variable illuminator illuminates a specimen from a plurality of N different incidence angles at different sample times. The optical element filters light issuing from the specimen. The radiation detector captures a plurality of variably-illuminated (perspective) low-resolution intensity images. The processor iteratively stitches together the variably-illuminated, low-resolution images of overlapping regions in Fourier space to recover a wide-field, high-resolution image. In certain embodiments, the FPI device may also correct for aberrations and digitally refocus the complex high-resolution image, which can digitally extend the depth of focus of the FPI system beyond the physical limitations of its optical element.

One embodiment provides a Fourier ptychographic imaging device comprising a variable illuminator for providing illumination to a specimen from a plurality of incidence angles, an optical element for filtering illumination issuing from the specimen, and a detector for acquiring a plurality of variably-illuminated, low-resolution intensity images of the specimen based on light filtered by the optical element. The Fourier ptychographic imaging device also comprises a processor for computationally reconstructing a high-resolution image of the specimen by iteratively updating overlapping regions in Fourier space with the variably-illuminated, low-resolution intensity images. In one case, the variable illuminator is a two-dimensional matrix of light elements (e.g., light-emitting diodes), each light element providing illumination from one of the plurality of incidence angles. In one example, each light element of the two-dimensional matrix comprises three quasi-monochromatic light sources and the processor computationally reconstructs high-resolution images corresponding to the three quasi-monochromatic light sources, and combines the high-resolution images to generate a color high-resolution image.

Another embodiment provides a method of Fourier ptychographic imaging. The method illuminates a specimen being imaged from a plurality of incidence angles using a variable illuminator and filters light issuing from (e.g., scattered by) the specimen using an optical element. The method also captures a plurality of variably-illuminated, low-resolution intensity images of the specimen using a detector. Also, the method computationally reconstructs a high-resolution image of the specimen by iteratively updating overlapping regions of variably-illuminated, low-resolution intensity images in Fourier space. In one case, the method initializes a current high-resolution image in Fourier space, filters an filtering an overlapping region of the current high-resolution image in Fourier space to generate a low-resolution image for an incidence angle of the plurality of incidence angles, replaces the intensity of the low-resolution image with an intensity measurement, and updates the overlapping region in Fourier space with the low-resolution image with measured intensity. In this case, the filtering, replacing, and updating steps may be performed for the plurality of incidence angles. In another case, the method divides each variably-illuminated, low-resolution intensity image into a plurality of variably-illuminated, low-resolution intensity tile images, recovers a high-resolution image for each tile by iteratively updating overlapping regions of variably-illuminated, low-resolution intensity tile images in Fourier space, and combines the high resolution images of the tiles to generate the high-resolution image of the specimen.

Another embodiment provides a method of Fourier ptychographic imaging that receives a plurality of variably-illuminated, low-resolution intensity images of a specimen and computationally reconstructs a high-resolution image of the specimen by iteratively updating overlapping regions of variably-illuminated, low-resolution intensity images in Fourier space. In one case, the method divides each variably-illuminated, low-resolution intensity image into a plurality of variably-illuminated, low-resolution intensity tile images, recovers a high-resolution image for each tile by iteratively updating overlapping regions of variably-illuminated, low-resolution intensity tile images in Fourier space, and combines the high-resolution images of the tiles. In another case, the method initializes a current high-resolution image in Fourier space, filters an overlapping region of the current high-resolution image in Fourier space to generate a low-resolution image for an incidence angle of the plurality of incidence angles, replaces intensity of the low-resolution image with an intensity measurement, and updates the overlapping region in Fourier space with the low-resolution image with measured intensity. In this case, the filtering, replacing, and updating steps may be performed for the plurality of incidence angles.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A is a schematic diagram of components of an FPI system, according to embodiments of the invention.

FIG. 1B is a schematic diagram of a side view of some components of the FPI device of FIG. 1A.

FIG. 2A is a schematic diagram of a FPI device comprising a variable illuminator in the form of a two-dimensional (10×10) matrix of 100 light elements, according to an embodiment of the invention.

FIG. 2B is a photograph of an FPI system with components in modular form, according to embodiments of the invention.

FIG. 2C is a photograph of one of the light elements of the variable illuminator the FPI device of FIG. 2B.

FIG. 3 is a schematic diagram of a side view of components of an FPI device, according to embodiments of the invention.

FIG. 4A is a schematic diagram of a side view of components of an FPI device, according to embodiments of the invention.

FIG. 4B is a schematic diagram of a side view of components of an FPI device, according to embodiments of the invention.

FIG. 5A includes a schematic representation of the measurement process (middle) and a recovery process (right-hand side) of an FPI method, according to embodiments of the invention.

FIGS. 5B(1), 5B(2), 5B(3), 5B(4), 5B(5), 5B(6), 5B(7), 5B(8), and 5B(9) are nine low-resolution measurements acquired by the FPI method introduced in FIG. 5A.

FIG. 5B(12) are regions updated in Fourier space associated with low-resolution measurements of FIGS. 5B(1), 5B(2), 5B(3), 5B(4), 5B(5), 5B(6), 5B(7), 5B(8), and 5B(9).

FIGS. 5B(10) and 5B(11) are the recovered high-resolution intensity and phase images resulting from the updating of FIG. 5B(12).

FIG. 6A is a flowchart of an FPI method performed by an FPI system, according to embodiments of the invention.

FIG. 6B is a flowchart of sub-steps of step 1500 of FIG. 6A, according to an embodiment of the invention.

FIGS. 6C and 6D are schematic diagrams of components of an FPI device with light elements in the form of an LED matrix, according to embodiments of the invention.

FIG. 6E is an illustration of steps of the FPI method described with reference to FIGS. 6A and 6B.

FIG. 6F is another illustration of steps of the FPI method described with reference to FIGS. 6A and 6B.

FIGS. 7A(1), 7A(2), 7A(3), 7A(4), and 7A(5) are images resulting from performing the FPI method of FIGS. 6A and 6B.

FIGS. 7B(1), 7B(2), 7B(3), 7B(4), 7B(5) and 7B(6) are images resulting from performing the FPI method of FIGS. 6A and 6B with different N numbers (N=5, 64, and 137) of incidence angles.

FIG. 8A is a flowchart of an FPI method with tile imaging, according to an embodiment of the invention.

FIG. 8B are images resulting from preforming an FPI method with tile imaging using image blending, according to an embodiment of the invention.

FIG. 9A is an FPI method with digital wavefront correction, according to an embodiment of the invention.

FIG. 9B is a schematic diagram of an FPI device implementing an FPI method with digital wavefront correction, according to an embodiment of the invention.

FIG. 10A are images from a numerical simulation of an FPI system 10 using an FPI method with and without digital refocusing for comparison, according to an embodiment of the invention.

FIGS. 10B(1)-(16) show experimental results from performing an FPI method with digital refocusing, according to an embodiment of the invention, using the FPI device 100(a) shown in FIG. 2B.

FIG. 10B(1) is a schematic diagram of the experimental setup of an FPI device performing an FPI method with digital refocusing, according to an embodiment.

FIGS. 10B(2)-(16) are images of the experimental results from performing the FPI method with digital refocusing according the experimental setup in FIG. 10B(1).

FIGS. 10C(1)-(7) include more detailed results from the experiment described with respect to FIGS. 10B(1)-10B(16).

FIGS. 10D(1)-(3) are images of experimental results from performing the FPI method with and without digital refocusing for comparison, according to an embodiment of the invention.

FIGS. 10E(1), 10E(2), 10E(3), 10E(4), 10E(5), 10E(6), 10E(7), and 10E(8) provides exemplary results from using an FPI system to perform an FPI method with digital refocusing, according to an embodiment of the invention, that corrects for chromatic aberration in a blood smear and in a pathology slide.

FIGS. 10F(1), 10F(2), 10F(3), 10F(4), 10F(5), 10F(6), and 10F(7) include experimental results from using an FPI system to perform an FPI method with digital autofocusing, according to an embodiment of the invention.

FIG. 11A(1) is a photograph illustrating the fields-of-view of a pathology slide for both 2× and 20× objective lenses of conventional microscopes, for comparison.

FIGS. 11A(2) and 11A(3) are images illustrating the numerical aperture for the 2× objective lens in a conventional microscope.

FIGS. 11A(4) and 11A(5) are images illustrating the numerical aperture for the 20× objective lens in a conventional microscope.

FIGS. 11A(6) and FIG. 11A(7) are color images showing the field-of-view and corresponding maximum NA of an FPI system, according to an embodiment of the invention.

FIGS. 11B(1)-(21) are images resulting from using a color imaging FPI system, according to an embodiment of the invention.

FIGS. 11C(1), 11C(2), 11C(3), 11C(4), 11C(5) and 11C(6) are images showing a comparison of image quality between the color imaging FPI system and different objective lenses, according to embodiments of the invention.

FIGS. 11D(1)-(10) are phase and color images from using an FPI method with the color imaging FPI system 10 both a pathology slide and blood smear, according to an embodiment of the invention.

FIG. 12 is a block diagram of subsystems that may be present in a FPI system 10, according to embodiments.

DETAILED DESCRIPTION OF THE INVENTION

Embodiments of the present invention will be described below with reference to the accompanying drawings. Although embodiments of FPI systems, devices, and methods may be described herein with respect to illumination with visible light radiation, these FPI systems, devices, and methods may also be used with other forms of radiation such as, for example, acoustic waves, Terahertz waves, microwaves, and X-rays.

Some embodiments include an FPI system comprising a variable illuminator, optical element, radiation detector, and a processor. The variable illuminator successively illuminates a specimen being imaged with plane waves at a plurality of N different incidence angles. The optical element filters light issuing from the specimen. The optical element may be, for example, an objective lens that accepts light issuing from the specimen based on its numerical aperture. In some cases, the optical element may be a low numerical aperture objective lens that provides a corresponding narrow acceptance angle and increased depth of field. The radiation detector detects light filtered by the optical element and captures a plurality of N low-resolution intensity images corresponding to the N different incidence angles. The processor iteratively stitches together overlapping low-resolution intensity images in Fourier space to recover a wide-field, high-resolution image of the specimen. In certain embodiments, the FPI device can also digitally refocus the complex high-resolution image to accommodate for defocus and aberrations in its optical element, which may digitally extend the depth of focus of the FPI system beyond the physical limitations of its optical element.

In certain aspects, an FPI method, performed by the FPI system, comprises a measurement process, a recovery process, and an optional display process. During the measurement process, the specimen is successively illuminated from the plurality of N incidence angles and corresponding low-resolution intensity images are acquired. During the recovery process, one or more high-resolution, wide field-of-view images are recovered based on the low-resolution intensity measurements. During the optional display process, the recovered images and other output is provided to the user of the FPI system on a display.

I. Introduction to FPI Systems and Devices

Although embodiments of FPI devices and systems are described herein with respect to visible light radiation (illumination), other forms of radiation (e.g., X-ray) may be used in certain cases.

FIG. 1A is a schematic diagram of components of an FPI system 10, according to embodiments of the invention. The FPI system 10 comprises an FPI device 100 and a computing device 200 in electronic communication with FPI device 100. In certain embodiments, such as the one illustrated in FIG. 1A, a specimen 20 is provided to the FPI device 100 for imaging. The FPI device 100 comprises a variable illuminator 110 for providing variable illumination to the specimen 20, an optical element 130 for filtering illumination issuing from the specimen 20, and a radiation detector 140 for detecting intensity of illumination received. The computing device 200 comprises a processor 210 (e.g., a microprocessor), a computer readable medium (CRM) 220, and a display 230.

During a measurement process, the variable illuminator 110 provides illumination from a plurality of N incidence angles, (θ_xⁱ, θ_yⁱ), i=1 to N to the specimen 20. Illumination from variable illuminator 110 may be altered (e.g., blocked, reduced intensity, modified wavelength/phase, modified polarization, etc.) by specimen 20 provided to the FPI device 100. The optical element can receive light from the variable illuminator, for example, as issuing from the specimen 20 and can filter the light it receives. For example, the optical element 130 may be in the form of an objective lens, which accepts light within its acceptance angle to act as a filter. In some cases, the optical element 130 may be an objective lens having a low numerical aperture (e.g., NA of about 0.08) to provide a corresponding narrow acceptance angle and allow for an increased depth of field. The radiation detector 140 can receive the filtered light from the optical element 130 and can record an intensity distribution at the radiation detector 140 at N sample times, t_{i=1 to N}, to capture a plurality of N low-resolution two-dimensional intensity images of the specimen area.

In FIG. 1A, the processor 210 is in electronic communication with radiation detector 140 to receive signal(s) with the image data corresponding to N low-resolution intensity images of the specimen area, which may include an image of at least a portion of the specimen 20. During a recovery process, the processor 210 can iteratively “stitch” together low-resolution intensity images in Fourier space to recover a wide-field, high-resolution image. In certain embodiments, the processor 210 can also digitally refocus the high-resolution image to accommodate for any defocus of the specimen and/or chromatic aberrations in its optical element 130. This capability can digitally extend the depth of focus of the FPI system 10 beyond the physical limitations of optical element 130.

Processor 210 is in electronic communication with CRM 220 (e.g., memory) to be able to transmit signals with image data in order to store to and retrieve image data from the CRM 220. Processor 210 is shown in electronic communication with display 230 to be able to send image data and instructions to display the wide-field, high-resolution image of the specimen area and other output, for example, to a user of the FPI system 10. As shown by a dotted line, variable illuminator 110 may optionally be in electronic communication with processor 210 to send instructions for controlling variable illuminator 110. As used herein, electronic communication between components of FPI system 10 may be in wired or wireless form.

FIG. 1B is a schematic diagram of a side view of some components of the FPI device 100 of FIG. 1A. In FIG. 1B, the FPI device 100 comprises a variable illuminator 110 having an illuminator surface 111, an optical element 130, and a radiation detector 140 having a sensing surface 142. Although radiation detector 140 is shown at a distance away from optical element 130, radiation detector 140 may optionally be located at the optical element 130.

In certain embodiments, the FPI device comprises an in-focus plane 122 and a sample plane 124. An in-focus plane 122 can refer to the focal plane of the optical element of the corresponding FPI device. The FPI device includes an x-axis and a y-axis at the in-focus plane 122, and a z-axis orthogonal to the in-focus plane 122. The in-focus plane is defined at an x-y plane at z=0. A sample plane 124 can refer to the plane at which the FPI device may computationally reconstruct the high-resolution wide field-of-view image. The FPI device captures the low-resolution images at the in-focus plane 122. Generally, the sample plane 124 is parallel to the in-focus plane 122. In some embodiments, the sample plane 124 may be coincident to the in-focus plane 122. In an autofocus embodiment, the FPI system 10 may perform an FPI method that can determine the location of the specimen 20 to locate the sample plane 124 at the specimen 20 in order to focus the high-resolution wide field-of-view image at the specimen 20.

In FIG. 1B, the FPI device 100 includes an in-focus plane 122 at z=0 and a sample plane at z=z₀. The FPI device 100 includes an x-axis and a y-axis (not shown) in the in-focus plane 122, and a z-axis orthogonal to the in-focus plane 122. The FPI device 100 also includes a distance d between the variable illuminator 110 and the sample plane 124. In the illustrated example, specimen 20 has been located at a specimen surface 126 for imaging. In other embodiments, specimen 20 may be in other locations for imaging purposes.

In FIG. 1B, the FPI device 100 is shown at a particular sample time, t_i, in the measurement process. At sample time, t_i, variable illuminator 110 provides incident illumination at a wavevector k_xⁱ, k_yⁱ associated with an incidence angle of (θ_xⁱ, θ_yⁱ) at the sample plane 124. Since the illustration is a side view in an x-z plane, only the x-component θ_xⁱ of the incidence angle is shown.

In FIG. 1B, the optical element 130 receives and filters light issuing from specimen 20. Light filtered by the optical element 130 is received at the sensing surface 142 of the radiation detector 140. The radiation detector 140 senses the intensity distribution of the filtered light and captures a low-resolution intensity image of the specimen area. Although FPI device 100 is shown at a sample time, t_i, the FPI device 100 can operate during a plurality of N sample times, t_{i=1 to N} to capture N low-resolution two-dimensional intensity images associated with N incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N.

A variable illuminator can refer to a device that provides incident radiation from a plurality of N different incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N, in succession. Suitable values of N may range from 2 to 1000. In most embodiments, the variable illuminator includes a light element of one or more radiation sources providing illumination at a particular sample time. In most cases, each light element is approximated as providing plane wave illumination to the specimen 20 from a single incidence angle. For example, the incidence angle θ_xⁱ at reference point, P, in FIG. 2A may be the angle between a normal and a line between point, P and the illuminated light element 112, which is based on a distance d between the variable illuminator and the sample plane 124.

Although the radiation source or radiation sources are usually coherent radiation sources, incoherent radiation source(s) may also be used and computational corrections may be applied. In embodiments that use visible light radiation, each radiation source is a visible light source. Some examples of a source of visible light include an LCD pixel and a pixel of an LED display. In embodiments that use other forms of radiation, other sources of radiation may be used. For example, in embodiments that use X-ray radiation, the radiation source may comprise an X-ray tube and a metal target. As another example, in embodiments that use microwave radiation, the radiation source may comprise a vacuum tube. As another example, in embodiments that use acoustic radiation, the radiation source may be an acoustic actuator. As another example, in embodiments that use Terahertz radiation, the radiation source may be a Gunn diode. One skilled in the art would contemplate other sources of radiation.

In many embodiments, the properties (e.g., wavelength(s), frequency(ies), phase, amplitude, polarity, etc.) of the radiation provided by the variable illuminator at different incidence angles, (θ_xⁱ, θ_yⁱ), i=1 to N, is approximately uniform. In other embodiments, the properties may vary at the different incidence angles, for example, by providing n different wavelengths λ₁, . . . , λ_n during the measurement process. In one embodiment, the variable illuminator 110 may provide RGB illumination of three wavelengths λ₁, λ₂, and λ₃ corresponding to red, green, blue colors, respectively. In embodiments that use Terahertz radiation, the frequencies of the radiation provided by the variable illuminator 110 may be in the range of 0.3 to 3 THz. In embodiments that use microwave radiation, the frequencies of the radiation provided by the variable illuminator may be in the range of 100 MHz to 300 GHz. In embodiments that use X-ray radiation, the wavelengths of the radiation provided by the variable illuminator may be in the range of 0.01 nm to 10 nm. In embodiments that use acoustic radiation, the frequencies of the radiation provided by the variable illuminator may be in the range of 10 Hz to 100 MHz.

In some embodiments, the variable illuminator comprises a plurality of N stationary light elements at different spatial locations (e.g., variable illuminator 110(a) in FIG. 2A). These N stationary light elements illuminate at N sample times in succession to provide illumination from the plurality of N incidence angles, (θ_xⁱ, θ_yⁱ), i=1 to N. In other embodiments, the variable illuminator comprises a moving light element (e.g., variable illuminator 110(b) in FIG. 3). This moving light element moves relative to the optical element and radiation detector, which may be kept stationary. In these embodiments, the moving light element may be moved to a plurality of N different spatial locations using a mechanism such as a scanning mechanism. Based on this relative movement between the stationary components and light element to the N different spatial locations, the light element can provide illumination from the plurality of N incidence angles, (θ_xⁱ, θ_yⁱ), i=1 to N. In other embodiments, the variable illuminator comprises a stationary light element (e.g., variable illuminator 110(c) in FIG. 4A) and the other components of the FPI device are moved to N different spatial locations. Based on this relative movement between the stationary light element and the other components of the FPI device to the N different spatial locations, the light element can provide illumination from the plurality of N incidence angles, (θ_xⁱ, θ_yⁱ), i=1 to N.

In embodiments having a variable illuminator comprising a plurality of N stationary light elements, the light elements may be arranged in the form of a one-dimensional array, a two-dimensional matrix, a hexagonal array, or other suitable arrangement capable of providing the illumination from the plurality of incidence angles. Some examples of matrices of stationary light elements are an LCD or an LED matrix. The light elements are designed with the appropriate spacing and designed to illuminate as required to provide the plurality of incidence angles. In some embodiments, the variable illuminator may be in the form of a two-dimensional matrix having dimensions such as for example, 10×10, 32×32, 100×100, 50×10, 20×60, etc. As an illustration example, FIG. 2A is a schematic diagram of a FPI device 100(a) comprising a variable illuminator 110(a) in the form of a two-dimensional (10×10) matrix of 100 stationary light elements 112, according to an embodiment of the invention.

In embodiments having a variable illuminator comprising a moving light element, the moving light element may be moved to a plurality of N positions. The spatial locations of these N positions may be in the form of a one-dimensional array, a two-dimensional matrix, a hexagonal array, or other suitable arrangement capable of providing the illumination from the plurality of incidence angles. Some examples of matrix dimensions may be 10×10, 32×32, 100×100, 50×10, 20×60, etc.

The variable illuminator provides radiation incident to the specimen 20 at a plurality of incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N. In one embodiment, the difference between two neighboring incidence angles in the plurality of incidence angles has a value in the range between 10% and 90% of the acceptance angle defined by the numerical aperture of the optical element in the form of an objective lens. In one embodiment, the difference between two adjacent incidence angles in the plurality of incidence angles has a value in the range between 33% and 66% of the acceptance angle defined by the numerical aperture of the optical element in the form of an objective lens. In one embodiment, the difference between two adjacent incidence angles in the plurality of incidence angles has a value that is less than 76% of the acceptance angle defined by the numerical aperture of the optical element in the form of an objective lens. In one embodiment, the difference between adjacent incidence angles in the plurality of incidence angles is about ⅓ of the acceptance angle defined by the numerical aperture of the optical element in the form of an objective lens. In one embodiment, the range of incidence angles, defined by a difference between the largest and smallest incidence angles, may be about equal to the effective numerical aperture consistent with the spatial resolution of the final full field-of-view high-resolution image.

The light elements of the variable illuminator are illuminated in an order defined by illumination instructions. In one embodiment, the illumination instructions determine the order of illuminating light elements in the form of a two-dimensional matrix of light elements. In this embodiment, the illumination instructions may first define a center light element. The illumination instructions may then instruct to illuminate the center light element (e.g., LED) first, then illuminate the 8 light elements surrounding the center light element going counterclockwise, then illuminate the 16 light elements surrounding the previous light element going counterclockwise, and so on until the N light elements have been illuminated from the plurality of N incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N. In another embodiment, the illumination instructions determine another order of illuminating light elements in the form of a two-dimensional matrix of light elements. In this embodiment, the variable illumination instructions may define a light element in the matrix that is closest to the specimen. The illumination instructions may then instruct to illuminate the light element closest to the specimen, and then illuminate the light element next closest to the specimen, and then illuminate the light element next closest, and so on until the N light elements have been illuminated from the plurality of N incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N.

In certain embodiments, the FPI device can image at least a portion of specimen 20 provided to the FPI device for imaging. In certain cases, the specimen 20 may comprise one or more objects. Each object may be a biological or inorganic entity. Examples of biological entities include whole cells, cell components, microorganisms such as bacteria or viruses, cell components such as proteins, thin tissue sections, etc. In some cases, the specimen 20 may be provided to the FPI device in a medium such as a liquid. In most cases, the specimen 20 is a stationary specimen. The specimen 20 is provided to the FPI device at a location capable of receiving illumination from the variable illuminator and so that light issuing from the specimen 20 is received by the optical element.

In certain embodiments, the FPI system 10 may comprise a receptacle for the specimen 20 with a specimen surface 126 for receiving a specimen 20. The specimen surface 126 may be part of a component of the FPI device 100, such as, for example, a surface of the variable illuminator 110. Alternatively, the specimen surface 126 may be a separate component from the FPI device 100 and/or FPI system 10. For example, the specimen surface 126 may a surface of a slide or a dish. This receptacle and specimen surface 126 may not be included in other embodiments.

In certain embodiments, one or more of the full field-of-view low-resolution images captured by the FPI device may be divided into one or more low-resolution tile images. In these cases, the processor can computationally reconstruct a high-resolution image for each tile independently, and then combine the tile images to generate the full field-of-view high-resolution image. This capability of independent processing of the tile images allows for parallel computing. In these embodiments, each tile may be represented by a two-dimensional area. The FPI system 10 uses an FPI method that assumes plane wave illumination over the area of each tile. In rectilinear spatial coordinates, each tile may be represented as a rectangular area (e.g., square area). In polar spatial coordinates, each tile may be a circular area or an oval area. In rectilinear spatial coordinates, the full field-of view low resolution image may be divided up into a two-dimensional matrix of tiles. In some embodiments, the dimensions of a two-dimensional square matrix of tiles may be in powers of two when expressed in number of pixels of the radiation sensor such as, for example, a 256 by 256 matrix, a 64×64 matrix, etc. In most cases, the tiles in the matrix will have approximately the same dimensions.

The FPI device also comprises an optical element that acts a low-pass filter. For example, the optical element may be an objective lens that only accepts light within a range of incidence angles based on its numerical aperture (NA). In many embodiments, the optical element is in the form of a low NA objective lens to provide narrow acceptance angle and high depth of field. In one embodiment, the optical element is a low NA objective lens has a low NA of about 0.08. In another embodiment, the optical element is a low NA objective lens has a low NA in the range between about 0.01 and about 0.1. In an embodiment of certain illustrated examples, the optical element is a 2× objective lens with an NA of about 0.08.

In embodiments that use X-ray radiation, an X-ray optical element may be needed, such as, for example, a grazing incidence mirror or zone plane. In embodiments that use acoustic radiation, a particular optical element may be needed such as, for example, an acoustic lens. In embodiments that use Terahertz radiation, a particular optical element may be needed such as, for example, a Teflon lens. In embodiments that use microwave radiation, a particular optical element may be needed such as, for example, a microwave lens antenna.

In certain embodiments, the FPI device has an initial depth of focus associated with the inherent depth of field of its optical element. A specimen provided to an FPI device of embodiments may be considered in focus when the sample plane is within the initial depth of focus of the optical element. Conversely, the specimen may be considered out-of-focus when the sample plane 124 is located outside of the initial depth of focus. Using an FPI method with digital refocusing of embodiments, the depth of focus of the FPI device may be extended beyond the limitations of the inherent depth of field of its optical element.

A radiation detector can refer to a device that can sense intensity of the radiation incident upon the radiation detector and can record spatial images based on the intensity pattern of the incident radiation. The radiation detector may record the images during a measurement process with a duration that includes at least the plurality of N sample times, t_{i=1 to N}. For an FPI device using visible light radiation, the radiation detector 140 may be, for example, in the form of a charge coupled device (CCD), a CMOS imaging sensor, an avalanche photo-diode (APD) array, a photo-diode (PD) array, or a photomultiplier tube (PMT) array. For an FPI device using THz radiation, the radiation detector may be, for example, an imaging bolometer. For an FPI device using microwave radiation, the radiation detector may be, for example, an antenna. For an FPI device using X-ray radiation, the radiation detector may be, for example, an x-ray sensitive CCD. For an FPI device using acoustic radiation, the radiation detector may be, for example, a piezoelectric transducer array. These radiation detectors and others are commercially available. In certain color imaging embodiments, the radiation detector may be a color detector e.g. an RGB detector. In other color imaging embodiments, the radiation detector need not be a color detector. In certain embodiments, the radiation detector may be a monochromatic detector.

A sample time can refer to a time that the radiation detector can capture a low-resolution image. In many embodiments, each sample time t_i, and associated captured low-resolution intensity image correspond to a particular incidence angle (θ_xⁱ, θ_yⁱ). The radiation detector may capture any suitable number N (e.g., 10, 20, 30, 50, 100, 1000, 10000, etc.) of low-resolution intensity images. The radiation detector may have a sampling rate or may have different sampling rates at which the radiation detector samples data. In some cases, sampling may be at a constant rate. In other cases, sampling may be at a variable rate. Some suitable examples of sample rates range from 0.1 to 1000 frames per second.

The radiation detector may have discrete radiation detecting elements (e.g., pixels). The radiation detecting elements may be of any suitable size (e.g., 1-10 microns) and any suitable shape (e.g., circular, rectangular, square, etc.). For example, a CMOS or CCD element may be 1-10 microns and an APD or PMT light detecting element may be as large as 1-4 mm. In one embodiment, the radiation detecting element is a square pixel having a size of 5.5 um.

The radiation detector can determine intensity image data related to captured low-resolution images. For example, the image data may include an intensity distribution. Image data may also include the sample time that the light was captured, or other information related to the intensity image.

Fourier space can refer to the mathematical space spanned by wavevectors k_x and k_y, being the coordinate space in which the two-dimensional Fourier transforms of the spatial images created by the FPI reside. Fourier space also can refer to the mathematical space spanned by wavevectors k_x and k_y in which the two-dimensional Fourier transforms of the spatial images collected by the radiation sensor reside.

Each of the low-resolution images captured by the radiation detector is associated with a region in Fourier space. This region in Fourier space can be defined by an approximated optical transfer function of the optical element and also by the incidence angle. If the optical element is an objective lens, for example, the low-resolution image in Fourier space may be the circular region defined by the approximated optical transfer function of the objective lens as a circular pupil with a radius of NA*k₀, where k₀ equals 2π/λ (the wave number in vacuum). In this example, the region is centered about the wave vector (k_xi, k_yi) associated with the corresponding incidence angle. In this example, the plurality of N low-resolution images are associated with a plurality of N regions centered about the plurality of N incidence angles in Fourier space.

In Fourier space, neighboring regions may share an overlapping area over which they sample the same Fourier domain data. The overlapping area between adjacent regions in Fourier space may be determined based on the values of the corresponding incidence angles. In most embodiments, the N incidence angles are designed so that the neighboring regions in Fourier space overlap by a certain amount of overlapping area. For example, the values of the N incidence angles may be designed to generate a certain amount of overlapping area for faster convergence to a high-resolution solution in the recovery process. In one embodiment, the overlapping area between neighboring regions may have an area that is in the range of 2% to 99.5% of the area of one of the regions. In another embodiment, the overlapping area between neighboring regions may have an area that is in the range of 65% to 75% the area of one of the regions. In another embodiment, the overlapping area between neighboring regions may have an area that is about 65% of the area of one of the regions.

The FPI system 10 of FIG. 1A also includes a computing device 200 that comprises a processor 210 (e.g., microprocessor), a CRM 220 (e.g., memory), and a display 230. The image display 230 and the CRM 220 are communicatively coupled to the processor 210. In other embodiments, the computing device 200 can be a separate device from the FPI system 10. The computing device 200 can be in various forms such as, for example, a smartphone, laptop, desktop, tablet, etc. Various forms of computing devices would be contemplated by one skilled in the art.

The processor 210 (e.g., microprocessor) may execute instructions stored on the CRM 220 to perform one or more functions of the FPI system 10. For example, the processor 210 may execute instructions to perform one or more steps of the recovery process of the FPI method. As another example, the processor 210 may execute illumination instructions for illuminating light elements of the variable illuminator. As another example, the processor 210 may execute instructions stored on the CRM 220 to perform one or more other functions of the FPI system 10 such as, for example, 1) interpreting image data from the plurality of low-resolution images, 2) generating a high-resolution image from the image data, and 3) displaying one or more images or other output from the FPI method on the display 230.

The CRM (e.g., memory) 220 can store instructions for performing some of the functions of the FPI system 10. The instructions are executable by the processor 210 or other processing components of the FPI system 10. The CRM 220 can also store the low-resolution and high-resolution image, and other data produced by the FPI system 10.

The FPI system 10 also includes a display 230 in electronic communication with the processor 210 to receive data (e.g., image data) and provide output data (e.g., images) to an operator of the FPI system 10. The image display 230 may be a color display or a black and white display. In addition, the display 230 may be a two-dimensional display or a three-dimensional display. In one embodiment, the display 230 may be capable of displaying multiple views.

Modifications, additions, or omissions may be made to FPI system 10 or FPI device 100 without departing from the scope of the disclosure. In addition, the components of FPI system 10 or the FPI device 100 may be integrated or separated according to particular needs. For example, the computing device 200 or components thereof may be integrated into the FPI device 100. In some embodiments, the processor 210 or other suitable processor may be part of the FPI device 100. In some cases, the processor 210 may be integrated into the radiation detector 140 so that the radiation detector 140 performs the functions of the processor 210. As another example, the CRM 220 and/or display 230 may be omitted from the FPI system 10 in certain cases.

II. FPI Device Configurations

In certain embodiments, FPI devices (e.g., FPI device 100(a) of FIG. 2A and FPI device 100(b) of FIG. 3) may be configured for use with particular types of radiation. For example, FPI device 100(a) of FIG. 2A may be particularly suitable for use with visible light, Terahertz, and/or microwave radiation. As another example, FPI device 100(c) of FIG. 4A may be particularly suitable for use with X-ray radiation.

FIG. 2A is a schematic diagram of a side view of components of an FPI device 100(a), according to embodiments of the invention. FPI device 100(a) comprises a variable illuminator 110(a) comprising a plurality of N stationary light elements, arranged in a two-dimensional matrix format. In the illustrated example, the i^th light element 112 provides illumination from an incidence angle (θ_xⁱ, θ_yⁱ). Although FIG. 2A shows the variable illuminator 110(a) having a 10×10 matrix of light elements 112, other dimensions can be used in other embodiments. In addition, although FIG. 2A shows equally spaced light elements 112 other spacings may be used in other embodiments. Variable illuminator 110(a) also comprises an x′-axis, y′-axis (not shown), and a z′-axis. As shown, the stationary light elements 112 extend in the x′-direction and the y′-direction.

FPI device 100(a) further comprises an optical element 130(a) (e.g., objective lens) and a radiation detector 140(a) having a sensing surface 142. Although radiation detector 140 is shown at a distance away from optical element 130(a), radiation detector 140 may optionally be located at the optical element 130(a). The FPI device 100(a) also includes an in-focus plane 122 at z=0 and a sample plane 124 at z=z₀. The FPI device 100(a) includes an x-axis and a y-axis (not shown) at the in-focus plane 122, and a z-axis orthogonal to the in-focus plane 122. The FPI device 100(a) also includes a distance d between the variable illuminator 110(a) and the sample plane 124. In the illustrated example, specimen 20 has been located at a specimen surface 126 for imaging. In other embodiments, specimen 20 may be in other locations for imaging purposes.

In FIG. 2A, the FPI device 100(a) is shown at a particular sample time, t_i, in the measurement process. At sample time, t_i, i^th light element 112 provides incident illumination at a wavevector k_xⁱ, k_yⁱ associated with an incidence angle of (θ_xⁱ, θ_yⁱ). The optical element 130(a) receives and filters light issuing from specimen 20. Light filtered by the optical element 130(a) is received at the sensing surface 142 of the radiation detector 140(a). The radiation detector 140(a) senses the intensity distribution of the filtered light and captures a low-resolution intensity image. Although FPI device 100(a) is shown at a single sample time, t_i, the FPI device 100(a) can operate at a plurality of N sample times, t_i=t_{i=1 to N}, associated with N incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N to capture N low-resolution two-dimensional intensity images.

In certain embodiments, components of an FPI system 10 may be in modular form to be placed in communication with components of a conventional microscope or other conventional imaging device to transform the conventional device into an FPI system 10. FIG. 2B is a photograph of an FPI system 10 with components in modular form, according to an embodiment of the invention. FPI system 10 comprises an FPI device 100(a). In the top photograph, the FPI device 100(a) comprises modular components that have been placed in communication with components of an Olympus® BX 41 microscope to transform these components of a conventional microscope into an FPI system 10. As an example of this modular aspect, the FPI device 100(a) includes a programmable two-dimensional LED matrix that has been placed under the specimen stage for illuminations. The programmable two-dimensional LED matrix comprises the plurality of light elements 112. FIG. 2C is a photograph of one of the light elements 112 of the variable illuminator 110(a) the FPI device 100(a) of FIG. 2B. This light element 112 is an LED that can provide red, green, and blue illuminations. As another example of the modular aspect of the illustrated example, the FPI device 110(a) in FIG. 2B comprises a radiation detector 140(a) in the form of a CCD camera. In FIG. 2B, the FPI device 100(a) further comprises an optical element 130(a) in the form of a 2×, 0.08 NA objective lens with an Olympus® BX 41 microscope. The field number of the 2× objective lens is 26.5. The field-of-view of the FPI device 100(a) at the sample plane is 13.25 mm in diameter. A processor 210 may be in electronic communication with the variable illuminator 110(a) and/or to radiation detector 140(a) through the wires 201.

In FIG. 2B, a specimen 20 has been provided to the FPI device 100(a) on a slide 202. Using red, green, and blue illuminations from the light elements 112, in this case LEDs, of the variable illuminator 110(a), a radiation detector of the FPI device 100 can acquire red, green, and blue low-resolution intensity images during a measurement process. The computing device 200 can computationally reconstruct a high-resolution and wide field-of-view color image of the specimen area by iteratively combining low-resolution measurements in Fourier space. In one case, the processor 210 may computationally reconstruct high-resolution and wide field-of-view red, green, and blue images, and then combine the images to generate a color image.

The FPI device 110(a) does not require a scanning mechanism for variable illumination. Other embodiments may include a scanning mechanism. For example, the FPI device 110(b) in FIG. 3 has a mechanism 150 that may be a scanning mechanism. As another example, the FPI device 110(c) in FIG. 4A has a mechanism 160 that may be a scanning mechanism.

FIG. 3 is a schematic diagram of a side view of components of an FPI device 100(b), according to an embodiment of the invention. FPI device 100(b) comprises a variable illuminator 110(b) comprising a light element 112 that is moved (e.g., scanned) in the x′-direction (direction on the x′-axis) and y′-direction (direction on the y′-axis) to a plurality of N locations. Variable illuminator 110(b) also comprises an x′-axis, y′-axis, and z′-axis. In the illustration, the light element 112 has moved from a normal incidence position (θ_xⁱ=0, θ_yⁱ=0) in the x′-direction to a position that provides illumination at (θ_xⁱ=−h, θ_yⁱ=0). The light element 112 is moved using a mechanism 150 (e.g., a raster scanner).

FPI device 100(b) further comprises an optical element 130(b) and a radiation detector 140(b) having a sensing surface 142. Although radiation detector 140(b) is shown at a distance away from optical element 130(b), radiation detector 140(b) may optionally be located at the optical element 130(b). The FPI device 100(b) also includes an in-focus plane 122 at z=0 and a sample plane 124 at z=z₀. The FPI device 100(b) includes an x-axis and a y-axis (not shown) at the in-focus plane 122, and a z-axis orthogonal to the in-focus plane 122. The FPI device 100(b) also includes a distance d between the variable illuminator 110(b) and the sample plane 124. In the illustrated example, specimen 20 has been located at a specimen surface 126 for imaging. In other embodiments, specimen 20 may be in other locations for imaging purposes.

In FIG. 3, the light element 112 is shown providing illumination at sample time, t_i in the measurement process. The optical element 130(b) filters light it receives. Light filtered by the optical element 130(b) is received at the sensing surface 142 of the radiation detector 140(b). The radiation detector 140(b) senses the intensity distribution of the filtered light and captures a low-resolution intensity image of the specimen area. Although FPI device 100(b) is shown at a single sample time, t_i, the FPI device 100(b) can operate at a plurality of N sample times, t_{i=1 to N}, associated with N incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N to capture N low-resolution two-dimensional intensity images. In embodiments where the FPI device 100(b) shown in FIG. 3 is used with X-ray radiation, the light element 112 includes an X-ray source.

FIG. 4A is a schematic diagram of a side view of components of an FPI device 100(c), according to embodiments of the invention. FPI device 100(c) comprises a variable illuminator 110(c) with a stationary light element 112, an optical element 130(c), a radiation detector 140(c) having a sensing surface 142, and a mechanism 160 (e.g., scanning mechanism). In the illustrated example, specimen 20 has been provided to the FPI device 100(c) for imaging.

In FIG. 4A, the mechanism 160 moves an assembly 170 comprising the optical element 130(c), a radiation detector 140(b) and specimen 20 relative to the stationary light element 112 to provide illumination from a plurality of N incidence angles. The mechanism 160 may translate and/or rotate the assembly 170. For example, the assembly 170 may mounted on a goniometer sate that would allow the assembly to be rotated as a whole relative to the light element 112. The variable illuminator 110(c) also comprises an x′-axis, y′-axis, and z′-axis.

Although radiation detector 140(c) is shown at a distance away from optical element 130(c), radiation detector 140(c) may optionally be located at the optical element 130(c). The FPI device 100(c) also includes an in-focus plane 122 at z=0 and a sample plane 124 at z=z₀. The FPI device 100(c) includes an x-axis and a y-axis (not shown) at the in-focus plane 122, and a z-axis orthogonal to the in-focus plane 122. The FPI device 100(c) also includes a distance d between the variable illuminator 110(c) and the sample plane 124.

In FIG. 4A, the light element 112 is shown providing illumination at sample time, t_i in the measurement process. The optical element 130(c) receives and filters light issuing from specimen 20. Light filtered by the optical element 130(c) is received at the sensing surface 142 of the radiation detector 140(c). The radiation detector 140(c) senses the intensity distribution of the filtered light and captures a low-resolution intensity image of the area. Although FPI device 100(c) is shown at a single sample time, t_i, the FPI device 100(c) can operate at a plurality of N sample times, t_{i=1 to N}, associated with N incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N to capture N low-resolution two-dimensional intensity images.

FIG. 4B is a schematic diagram of a side view of components of an FPI device 100(d), according to embodiments of the invention. FPI device 100(d) comprises a variable illuminator 110(d) with a light element 112 that is moved by rotating it, an optical element 130(b), and a radiation detector 140(b) having a sensing surface 142. Although not shown, a mechanism may also be included to rotate the light element 112. In the illustrated example, specimen 20 has been provided to the FPI device 100(d) for imaging. In some cases, the light element 112 may be a laser.

In FIG. 4B, the light element 112 is moved by rotating it, which provides illumination at (θ_xⁱ, θ_yⁱ). In FIG. 4B, the light element 112 is shown providing illumination at sample time, t_i in the measurement process. The optical element 130(b) receives and filters light issuing from specimen 20. Light filtered by the optical element 130(b) is received at the sensing surface 142 of the radiation detector 140(b). The radiation detector 140(b) senses the intensity distribution of the filtered light and captures a low-resolution intensity image of the area. Although FPI device 100(d) is shown at a single sample time, t_i, the FPI device 100(d) can operate at a plurality of N sample times, t_{i=1 to N}, associated with N incidence angles (θ_xⁱ, θ_yⁱ), i=1 to N to capture N low-resolution two-dimensional intensity images.

III. Exemplary FPI Methods

In embodiments, an FPI method comprises a measurement process, a recovery process, and an optional display process. In the measurement process, the specimen is illuminated from a plurality of incidence angles using the variable illuminator, the optical element filters light issuing from the specimen, and the radiation detector captures a plurality of low resolution intensity images based on the filtered light. In the recovery process, the intensity of each low resolution image obtained by inverse Fourier transformation and filtering of the high resolution reconstruction in Fourier space is replaced by the low-resolution intensity measurement and the corresponding region of the high-resolution reconstruction in Fourier space is iteratively updated. In the recovery process, the high-resolution image of the specimen may be computationally reconstructed based on a plurality of N low-resolution intensity images. In the optional display process, images and other output is provided to a display 220.

FIG. 5A includes a schematic representation of the measurement process (middle) and a recovery process (right-hand side) of an FPI method, according to embodiments of the invention. During the measurement process, a specimen is illuminated from different incidence angles, and low-resolution intensity images corresponding to these incidence angles are acquired. The acquisition of multiple low-resolution intensity images is represented by the arrangement of images in the middle section of FIG. 5A. During the recovery process, one or more high-resolution, wide field-of-view images are recovered based on the low-resolution intensity measurements from the measurement process. The recovery process is represented by the two images at the right-hand side of FIG. 5A, where both high-resolution intensity and phase image data are recovered based on the low-resolution intensity measurements. FIGS. 5B(1), 5B(2), 5B(3), 5B(4), 5B(5), 5B(6), 5B(7), 5B(8), and 5B(9) are nine low-resolution measurements of the 137 measurements acquired by the FPI method introduced in FIG. 5A. The corresponding regions associated with the low-resolution images in Fourier space are shown in FIG. 5B(12). During the recovery process, these regions are updated in Fourier space to reconstruct a full FOV high-resolution complex image. Recovered high-resolution intensity and phase images are shown at the right-hand side of FIG. 5A and also shown in FIGS. 5B(10) and 5B(11).

In certain embodiments, the FPI method may alternate between two working domains: the spatial (x-y) domain and Fourier (k_x-k_y) domain, where k represents the wavenumber.

In certain embodiments, the FPI method may use the concept of oblique incidence, which provides that illuminating the specimen 20 (e.g., a thin specimen) by an oblique plane wave with a wavevector (k_xⁱ, k_yⁱ) is equivalent to shifting the center of the image spectrum by (k_x,k_y) in the Fourier domain. According to oblique incidence, the low-resolution image in the Fourier domain is shifted from normal incidence by (k_x, k_y), which corresponds to the incidence angle applied by the variable illuminator.

In certain embodiments, the FPI method may provide that the filtering function of the optical element (i.e. the coherent optical transfer function) in Fourier space is a circular pupil with a radius of NA×k₀, where k₀=2π/λ is the wave number in vacuum. That is, the FPI method may update in Fourier space circular regions defined by this filtering function of the optical element. In these embodiments, the FPI method uses this filtering function to omit image data outside this region.

In certain embodiments, the specimen 20 may be placed at the sample plane 124 at z=z₀ where the in-focus plane 122 of the optical element is located at position z=0. In other words, the image captured is not the specimen image at the specimen profile itself; it is the specimen profile propagated by −z₀ distance at the in-focus plane of the optical element. In these embodiments, the FPI method may digitally re-focus the specimen 20 propagating the image data by the z₀ distance back to the sample plane 124, without having to mechanically move the specimen in the z-direction. These propagating step(s) can be performed by multiplying a phase factor in Fourier space. These steps can extend the imaging depth of focus of the FPI system 10 and correct the chromatic aberration of the optical element.

FIG. 6A is a flowchart of an FPI method performed by an FPI system 10, according to embodiments of the invention. The FPI method comprises a measurement process (steps 1100, 1200, and 1300), a recovery process (steps 1400 and 1500), and an optional display process (step 1600). In illustrated examples of FPI methods and their associated description, subscript “h” refers to high-resolution, subscript “l” refers to low resolution, subscript “f” refers to focused position, subscript “m” refers to measured, and subscript “s” refers to sampled.

At step 1100, the variable illuminator provides illumination to a specimen area from a plurality of N incidence angles, (θ_xⁱ, θ_yⁱ), i=1 . . . N, at N sample times. In most cases, the recovery process assumes plane wave illumination. The variable illuminator may provide illumination according to illumination instructions that define the order of the illumination angles. The wave vector in x and y directions can be denoted as k_xi and k_yi. In certain cases, the variable illuminator may provide illumination of different wavelengths at different sample times. For example, the variable illuminator 110 may provide RGB illumination of three wavelengths λ₁, λ₂, and λ₃ corresponding to red, green, blue colors, respectively, for a color imaging embodiment.

At step 1200, the optical element (e.g., a low-NA microscope objective lens) filters light it receives. For example, the optical element can filter light issuing from the specimen 20. The optical element may be an objective lens that filters light by accepting light within a range of incidence angles according to its numerical aperture (NA). In some cases, the optical element may be a low-NA objective lens (e.g., a 2×, 0.08 NA objective lens) of a conventional microscope.

At step 1300, the radiation detector receives a projection of the filtered light from the optical element and captures a snapshot intensity distribution measurement at each of the N sample times, t_{i=1 to N} to acquire a plurality of N low-resolution intensity images. Each low resolution intensity image sampled by the radiation detector is associated with a region in Fourier space. In many embodiments, the variable illuminator provides illumination from certain incidence angles at step 1100 to generate overlapping areas between neighboring regions in Fourier space. In one embodiment, the variable illuminator provides illumination to provide an overlapping area between neighboring regions of 2% to 99.5% of the area of one of the regions. In another embodiment, the variable illuminator provides illumination to provide an overlapping area between neighboring regions of 65% to 75% of the area of one of the regions. In one embodiment, the variable illuminator provides illumination to provide an overlapping area between neighboring regions of about 65% of the area of one of the regions.

In steps 1400 and 1500, a high-resolution image of the specimen area may be computationally reconstructed from the plurality of N low-resolution intensity distribution measurements, I_lm(kⁱ_x, k_yⁱ) (indexed by their illumination wavevector, k_xⁱ, k_yⁱ, with i=1, 2, . . . N) captured at step 1300.

At step 1400, a high-resolution image: √{square root over (I_h)}e^iφh is initialized in the spatial domain, and a Fourier transform is applied to the initial value to obtain an initialized Fourier transformed image Ĩ_h. The initialized high-resolution solution may be an initial guess. This initial guess may be determined based on the assumption that the specimen is located at the out-of-focus plane z=z₀. In some cases, the initial guess may be determined as a random complex matrix (for both intensity and phase). In other cases, the initial guess may be determined as an interpolation of the low-resolution intensity measurement with a random phase. An example of an initial guess is φ=0 and I_h interpolated from any low-resolution image of the specimen area. Another example of an initial guess is a constant value. The Fourier transform of the initial guess can be a broad spectrum in the Fourier domain.

At step 1500, the high-resolution image of the specimen area is computationally reconstructed by iteratively combining low-resolution intensity measurements in Fourier space using the processor 210 of the FPI system 10.

At optional step 1600, the display 230 may receive image data such as the high-resolution image data to √{square root over (I_h)}e^iφh and/or other data from the processor 210, and display the data on the display 230.

FIG. 6B is a flowchart of sub-steps of step 1500 of FIG. 6A, according to an embodiment of the invention. In this illustrated example, step 1500 comprises step 1510, step 1530, step 1550, step 1560, step 1570, step 1580, and step 1590. Step 1500 may optionally comprise steps 1520 and 1540. Optional steps 1520 and 1540 may be performed if the specimen 20 is out-of-focus by the amount of z₀.

At step 1510, the processor 210 performs low-pass filtering of the high-resolution image √{square root over (I_h)}e^iφh in the Fourier domain to generate a low-resolution image √{square root over (I_l)}e^iφl for a particular plane wave incidence angle (θ_xⁱ, θ_yⁱ) with a wave vector (k_xⁱ, k_yⁱ). The Fourier transform of the high-resolution image is Ĩ_h and the Fourier transform of the low-resolution image for a particular plane wave incidence angle is Ĩ_l. In the Fourier domain, the FPI method filters a low-pass region from the spectrum Ĩ_h of the high-resolution image √{square root over (I_h)}e^iφh. In cases with an optical element in the form of an objective lens, this region is a circular aperture with a radius of NA*k₀, where k₀ equals 2π/λ (the wave number in vacuum), given by the coherent transfer function of an objective lens. In Fourier space, the location of the region corresponds to the incidence angle. For an oblique plane wave incidence with a wave vector (k_xⁱ, k_yⁱ), the region is centered about a position (−k_xⁱ,−k_yⁱ) in the Fourier domain of √{square root over (I_h)}e^iφh.

At optional step 1520, using the processor 210, the low-resolution image, √{square root over (I_l)}e^iφl is propagated in the Fourier domain to the in-focus plane 122 at z=0 of the optical element 130 to determine the low-resolution image at the focused position: √{square root over (I_lf)}e^iφlf. In one embodiment, Step 1520 can be performed by Fourier transforming the low-resolution image √{square root over (I_l)}e^iφl, multiplying by a phase factor in the Fourier domain, and inverse Fourier transforming to obtain √{square root over (I_lf)}e^iφlf. In another embodiment, step 1520 can be performed by the mathematically equivalent operation of convolving the low-resolution image √{square root over (I_l)}e^iφl with the point-spread-function for the defocus. In another embodiment, step 1520 can be performed as an optional sub-step of step 1510 by multiplying by multiplying Ĩ_l by a phase factor in the Fourier domain before performing the inverse Fourier transform to produce √{square root over (I_lf)}e^iφlf. Optional step 1520 need not be included if the specimen 20 is located at the in-focus plane (z=0) of the optical element.

At step 1530, using the processor 210, the computed amplitude component √{square root over (I_lf)} of the low-resolution image at the in-focus plane, √{square root over (I_lf)}e^iφlf, is replaced with the square root of the low-resolution intensity measurement √{square root over (I_lfm)} measured by the radiation detector of the FPI device. This forms an updated low resolution target: √{square root over (I_lfm)}e^iφlf.

At optional step 1540, using the processor 210, the updated low-resolution image √{square root over (I_lfm)}e^iφlf may be back-propagated to the sample plane (z=z₀) to determine √{square root over (I_ls)}e^iφls. Optional step 1540 need not be included if the specimen is located at the in-focus plane of the optical element, that is, where z₀=0. In one embodiment, step 1540 can be performed by taking the Fourier transform of the updated low-resolution image √{square root over (I_lfm)}e^iφlf and multiplying in the Fourier space by a phase factor, and then inverse Fourier transforming it. In another embodiment, step 1540 can be performed by convolving the updated low-resolution image √{square root over (I_lfm)}e^iφlf with the point-spread-function of the defocus. In another embodiment, step 1540 can be performed as a sub-step of step 1550 by multiplying by a phase factor after performing the Fourier transform onto the updated target image.

At step 1550, using the processor 210, a Fourier transform is applied to the updated target image propagated to the sample plane: √{square root over (I_ls)}e^iφls, and this data is updated in the corresponding region of high-resolution solution √{square root over (I_h)}e^iφh in the Fourier space corresponding to the corresponding to the incidence wave vector (k_xⁱ, k_yⁱ).

At step 1560, the processor 210 determines whether steps 1510 through 1560 have been completed for all incidence angles. If steps 1510 through 1560 have not been completed for all incidence angles, steps 1510 through 1560 are repeated for the next incidence angle.

In most embodiments, the neighboring regions in Fourier space, which are iteratively updated for each incidence angle, overlap each other. In the overlapping area between updated overlapping regions, the FPI system 10 has multiple samplings over the same Fourier space. The incidence angles determine the area of the overlapping area. In one embodiment, the overlapping area between neighboring regions may have an area that is between 2% to 99.5% of the area of one of the neighboring regions. In another embodiment, the overlapping area between neighboring regions may have an area that is between 65% to 75% of the area of one of the neighboring regions. In another embodiment, the overlapping area between neighboring regions may have an area that is about 65% of the area of one of the neighboring regions. In certain embodiments, each overlapping region has the same area.

At step 1570, the processor 210 determines whether the high-resolution solution has converged (step 1570). For example, the processor 210 may determine whether the high-resolution solution may have converged to a self-consistent solution. In one case, the processor 210 compares the previous high-resolution solution of the previous iteration or initial guess to the present high-resolution solution, and if the difference is less than a certain value, the solution may have converged to a self-consistent solution. If processor 210 determines that the solution has not converged, then steps 1510 through 1570 are repeated. In one embodiment, steps 1510 through 1560 are repeated once. In other embodiments, steps 1510 through 1560 are repeated twice or more. If the solution has converged, the processor 210 transforms the converged solution in Fourier space to the spatial domain to recover a high-resolution image √{square root over (I_h)}e^iφh. If the processor 210 determines that the solution has converged at step 1570, then the process may proceed to optional step 1600.

FIG. 6E is an illustration of steps of the FPI method described with reference to FIGS. 6A and 6B. The left-hand-side image in FIG. 6E includes two regions 22(a) and 22(b) with circular low-pass filter shapes in Fourier space of the high-resolution solution, defined by the optical transfer function of a 2× objective lens 0.08 NA. Region 22(a) is based on a circular low-pass filter shape associated with normal plane wave incidence at the first incidence angle: θ_x=0; θ_y=0, i=1. Region 22(b) is based on a circular low-pass filter shape associated with an Nth plane wave incidence angle: θ_x=−21°; θ_y=22°; i=N. To perform low-pass filtering at each incidence angle, data outside the circular region in the Fourier domain is omitted, which results in a low-resolution image. The low-resolution image resulting from filtering based on oblique plane wave incidence angle: θ_x=−21; θ_y=22° is shown at the top right-hand-side of FIG. 6E. The low-resolution image resulting from filtering at this oblique plane wave incidence angle of θ_x=−21°; θ_y=22° is shown at the bottom right hand side. The wave vectors of the incidence angles in the x-direction and y-direction are denoted as k_xi and k_yi respectively. In this illustration, the dimensions of the regions 22(a) and 22(b) are defined by the optical transfer function of a 2× objective lens 0.08 NA based on approximating as circular pupil function with a radius of NA*k₀, where k₀ equals 2π/λ (the wave number in vacuum).

In FIG. 6E, the FPI method updates the data within region 22(a) of the high-resolution reconstruction corresponding to the normal incidence θ_x=0, θ_y=0, according to step 1550 of FIG. 6B. The region is updated with low-resolution image data having an updated intensity measurement I_ls (k_x¹, k_y^l) where k_x¹=0, k_y¹=0. In FIG. 6E, the FPI method also updated the data within region 22(b) of the high-resolution reconstruction corresponding to the n^th incidence angle θ_x=−21°; θ_y=22°; i=N, according to step 1550 of FIG. 6B. The region is updated with low-resolution image data having an updated intensity measurement I_ls(k_x^N, k_y^N).

FIG. 6F is another illustration of steps of the FPI method described with reference to FIGS. 6A and 6B. In this case, the specimen 20 is at the in-focus plane. The schematic drawing includes illustrations 1605, 1606, and 1607. Illustration 1605 includes a representative image of an initial Fourier spectrum resulting from step 1400 and the interpolated intensity. Illustration 1606 represents the iterative portion of the step 1500, which includes 1510, 1530, 1550 in this case. Illustration 1606 includes N iterations. Each of the regions 22(d), 22(e), and 22(f) are the circular low-pass filter shapes in Fourier space of the high-resolution solution. The dimensions of the regions 22(a) and 22(b) are defined by the optical transfer function of a 2× objective lens 0.08 NA based on approximating as circular pupil function with a radius of NA*k₀, where k₀ equals 2π/λ (the wave number in vacuum). At the first iteration (i=1), θ_x=0; θ_y=0, region 22(d) is updated at step 1550. At the N−1^th iteration (i=N−1), θ_x=−22; θ_y=19, region 22(e) is updated at step 1550. At the N^th iteration (i=N), θ_x=−22; θ_y=22, region 22(f) is updated at step 1550. Illustration 1606 includes the recovered phase image and the recovered intensity images.

In embodiments, the FPI method can include digital refocusing by including optional steps 1520 and 1540. The digital refocusing feature of steps 1520 and 1540 propagates the image from the in-focus plane 122 z=0 to the sample plane 124 at z=z₀. Digital refocusing may be needed when the specimen 20 is located at the sample plane 124 at z=z₀, while the in-focus plane 122 of the optical element (e.g., objective lens) is located at position z=0. In other words, digital refocusing may be needed when the specimen 20 is out-of-focus by the amount of z₀. FIGS. 6C and 6D are schematic diagrams of components of an FPI device 100(a) with light elements in the form of an LED matrix, according to an embodiment of the invention. In FIG. 6C, the specimen 20 is out-of-focus by an amount of −z₀, and optional steps 1520 and 1540 can be used to digitally refocus. In FIG. 6D, the specimen 20 is located at the in-focus position at the in-focus plane 122. In this case, optional steps 1520 and 1540 may not be needed.

In one embodiment, an FPI system 10 with the FPI device 100(a) shown in FIG. 2B was used to perform the FPI method of FIGS. 6A and 6B where the specimen 20 was located at the in-focus plane 122. In this example, the light elements were in the form of 137 red LEDs as the light sources for oblique illuminations. The corresponding maximum NA of the reconstructed image was ˜0.5 in the diagonal direction.

FIGS. 7A(1), 7A(2), 7A(3), 7A(4), and 7A(5) are images resulting from performing the FPI method of FIGS. 6A and 6B. These results show improved resolution. FIG. 7A(1) is a full field-of-view low-resolution intensity image of the specimen area. FIG. 7A(2) is a zoom-in-view of a selected portion of the low resolution image FIG. 7A(1) captured by the radiation detector, with a pixel size of 2.75 μm (5.5 μm divided by the magnification factor). FIG. 7A(3) is a zoom-in-view of the image in FIG. 7A(2). FIGS. 7A(4) and 7A(5) are computationally reconstructed high-resolution images, where the pixel size is 0.275 um.

In one embodiment, an FPI system 10 with the FPI device 100(a) shown in FIG. 2B was used to perform the FPI method of FIGS. 6A and 6B where the specimen 20 was located at the in-focus plane 122, with different numbers of LED light sources (i.e. different N number of incidence angles) and their corresponding high-resolution reconstructions with different maximum NAs covered in Fourier space with the reconstruction. In the first case, five low resolution images (frames) were acquired at five incidence angles (i=1 to 5). FIG. 7B(1) shows resulting high-resolution images from the five frame reconstruction. FIG. 7B(2) shows the total spectrum of the five frame reconstruction in Fourier space that spans a maximum synthetic NA of 0.13. In the second case, 64 low resolution images (frames) were acquired at 64 incidence angles (i=1 to 64). FIG. 7B(3) shows resulting high-resolution images from the 64 frame reconstruction. FIG. 7B(4) shows the spectrum of the 64 frame reconstruction in Fourier space that spans a maximum synthetic NA of 0.3. In the third case, 137 low resolution images were acquired at 137 incidence angles (i=1 to 137). FIG. 7B(5) shows resulting high-resolution images from the 137 frame reconstruction. FIG. 7B(6) shows the spectrum of the 137 frame reconstruction in Fourier space that spans a maximum synthetic NA of 0.5. Each small circle in FIGS. 7B(2), 7B(4), and 7B(6) represent the spectrum region corresponding to one low-resolution intensity measurement. These results show that the field-of-view can be decoupled from the resolution of the optics using the FPI method, and as such, achieves wide field-of-view and high-resolution at the same time.

A. FPI Methods with Tile Imaging

In some embodiments, an FPI method with tile imaging divides the captured low-resolution images into a plurality of low-resolution tile images, independently acquires a high-resolution image for each of the tiles, and then combines the high-resolution tile images to generate a full field-of-view high-resolution image. In some cases, the high-resolution tile images may be combined with an image blending process. An example of an image blending process is alpha blending which can be found in PCT publication WO1999053469, entitled “A system and method for performing blending using an over sampled buffer,” filed on Apr. 7, 1999, which is hereby incorporated by reference in its entirety. Since the high-resolution images of the tiles may be acquired independently, this FPI method may allow for parallel computing, which may reduce computation time, and may also reduce memory requirements. Moreover, the light from each light element may be accurately treated as a plane wave for each tile. The incident wavevector for each tile can be expressed as:

$\begin{array}{cc}\left(k_x^i,k_y^i\right)=\frac{2\pi }{\lambda }\left(\begin{array}{c}\frac{\left(x_c-x_i\right)}{\sqrt{{\left(x_c-x_i\right)}^2+{\left(y_c-y_i\right)}^2+h^2}},\\ \frac{\left(y_c-y_i\right)}{\sqrt{{\left(x_c-x_i\right)}^2+{\left(y_c-y_i\right)}^2+h^2}}\end{array}\right)& \left(\mathrm{Eqn}.1\right)\end{array}$

where (x_c,y_c) is the central position of each tile of the full field-of-view low-resolution image, (x_i,y_i) is the position of the i^th light element, and h is the distance between the illuminator and the specimen. Furthermore, this FPI method can assign a specific aberration-correcting pupil function to each tile in some cases.

In one embodiment, an FPI system 10 comprises a computing device 200 in the form of a personal computer having a processor 210 in the form of an Intel i7 CPU. In one experiment, this FPI system 10 performed an FPI method with tile imaging that divides the low-resolution images into tiles of 196×196 pixels. In one case, the processing time for each small image (converting 196 by 196 raw pixels to 1960 by 1960 pixels) is about 3.75 seconds using Matlab. In this case, the processing time for creating the full field-of-view gigapixel image is about 10 minutes when parallel computing using all 4 cores of the CPU. In other embodiments, an FPI system 10 can have a processor in the form of a GPU unit, which may reduce processing time by ten-fold. In other embodiments, the processor 210 may perform steps of the FPI method using another programming language, such as C++, which may reduce processing time. An example of this FPI system can be found in Zheng, G., Horstmeyer, R., and Yang, C., “Wide-field, high-resolution Fourier ptychographic microscopy,” Nature Photonics (July, 2013), which is hereby incorporated by reference in its entirety.

FIG. 8A is a flowchart of an FPI method with tile imaging, according to an embodiment of the invention. This FPI method can be performed by an FPI system 10. To take advantage of parallel processing capabilities, the FPI system 10 includes a processor 210 with parallel processing capabilities such as, for example, the GPU unit or a processor having multiple cores (i.e. independent central processing units). The FPI method comprises a measurement process (steps 1100, 1200, and 1300), a recovery process (steps 1350, 2400 (i-M), 2500(i-M), 2590), and an optional display process (step 1600). The measurements process (steps 1100, 1200, and 1300) and display process (step 1600) are described with reference to FIG. 6A.

At step 1350, the processor 210 divides the full field-of-view into a plurality of tiles such as, for example, a two-dimensional matrix of tiles. The dimensions of a two-dimensional square matrix of tiles may be in powers of two such as, for example, a 256 by 256 matrix, a 64×64 matrix, etc. In one example, the processor may divide up a full field of view of 5,280×4,380 pixels into tiles having an area of 150×150 pixels.

Next, the processor 210 initializes the high-resolution image: √{square root over (I_h)}e^iφh in the spatial domain for each tile (1 to M) independently using parallel computing (step 2400(1) . . . step 2400(M)). A Fourier transform is applied to the initial guess. In some cases, the initial guess may be determined as a random complex matrix (for both intensity and phase). In other cases, the initial guess may be determined as an interpolation of the low-resolution intensity measurement with a random phase. An example of an initial guess is φ=0 and I_hr of any low-resolution image of the specimen area. Another example of an initial guess is a constant value. The Fourier transform of the initial guess can be a broad spectrum in the Fourier domain.

At step 2500(1) . . . step 2500(M), the processor 210 computationally reconstructs a high-resolution image of each tile (1 to M) independently using parallel computing. The processor 210 computationally reconstructs the high-resolution image of each tile by iteratively combining low-resolution intensity images in Fourier space as described with reference to steps 1510, 1530, 1550, 1560, and 1570 shown in FIG. 6B, and described herein. Steps 1520 and 1540 may be included if the specimen is out of focus.

At step 2590, the processor 210 combines the high-resolution tile images into a full field-of view high-resolution image. In some cases, combining tile images comprises an imaging-blending process such as, for example, alpha blending.

FIG. 8B are images resulting from preforming an FPI method with tile imaging using image blending, according to an embodiment of the invention. In this example, the FPI method divided each low-resolution image of 5320-by-4370 pixels into 150-by-150 pixel sized tiles. On the left-hand-side, the images are of two neighboring high-resolution tile images. The image on the right-hand-side is a high-resolution image that combines the two high-resolution tile images using blending. In blending, the FPI method generates a blending region that replaces the edge regions of the adjacent tile images in the high-resolution full FOV image. In the illustration, the width of the blending region was 50 pixels. In other embodiments, the blending region may have other widths such as, for example, 6 pixels. With blending, there is no observable boundary in the blending region.

At optional step 1600, the image data of the recovered high-resolution two-dimensional image of the specimen area is displayed on the display 230.

In one embodiment, the FPI method with tile imaging may further comprise a procedure that accounts for differences in incident angles between different tiles based on the distance between the tiles and each light element.

B. FPI Methods with Digital Refocusing

Another limitation of conventional high-NA microscopes is the limited depth-of field. As an example, the depth-of-field of a conventional microscope with a 20× objective lens with 0.4 NA is about 5 μm. With a conventional microscope, the resolution degrades as the specimen moves away from the in-focus plane 122 due to its limited depth-of-field. To achieve optimal resolution using a conventional microscope, the operator needs to move the stage to mechanically bring the specimen back into focus. In this regard, a precise mechanical stage is needed in the conventional microscope to bring a specimen into the in-focus position with sub-micron accuracy.

In certain embodiments, an FPI system 10 implements an FPI method in which a specimen can be refocused digitally rather than mechanically. In these cases, the FPI method computationally refocuses the out-of-focus specimen 20 during the recovery process. Using digital refocusing, the FPI system 10 can expand its depth-of focus beyond the physical limitations of its optical element.

FIG. 10A are images from a numerical simulation of an FPI system 10 using an FPI method with and without digital refocusing for comparison, according to an embodiment of the invention. These images show that the high-resolution reconstruction resulting from using this FPI method is invariant to the specimen defocusing. In FIG. 10A, each row represents a different defocused amount in the z-direction of −150 μm, −50 μm, 50 μm, and 150 μm. Column 1 is the low-resolution image captured. Columns 2 and 3 are the recovered high-resolution intensity and phase profiles with digital refocusing. Columns 4 and 5 are the recovered high-resolution intensity and phase profiles without digital refocusing.

FIG. 10B(1) is a schematic diagram of the experimental setup of an FPI device 100(a) performing an FPI method with digital refocusing, according to an embodiment. FIGS. 10B(2)-(16) are images of the experimental results from performing the FPI method with digital refocusing according the experimental setup in FIG. 10B(1). In the experiment, the specimen was moved to different z positions with defocused distances ranging from −300 μm to +300 μm, as shown in the schematic diagram in FIG. 10B(1). Low-resolution images (corresponding to 137 different LEDs of an LED matrix) were acquired for all the defocused positions. Representative low resolution images are shown in FIGS. 10B(2), (3), (4), (5), and (6). The high-resolution profiles of the specimen were computationally reconstructed following the recovery process with steps 1520 and 1540 of FIG. 6B. These high-resolution profiles are shown in FIGS. 10B(7), (8), (9), (10), and (11). FIGS. 10C(1)-(7) include more detailed results from the experiment described with respect to FIGS. 10B(1)-10B(16). FIG. 10C(7) is a graph of line traces for the smallest features in FIGS. 10C(4), (5), and (6). As shown by the reconstructed images in FIGS. 10B and 10C, the FPI method with digital refocusing, can achieve resolution-invariant imaging depth of 0.3 mm.

FIGS. 10D(1)-(3) are images of experimental results from performing the FPI method with and without digital refocusing for comparison, according to an embodiment of the invention. In FIGS. 10D(2) and 10D(3), image reconstructions resulting from using an FPI method of embodiments are shown, with and without digital refocusing (e.g. steps 1520 and 1540), for comparison. FIG. 10D(1) is a low-resolution image of the raw data at z=−150 μm. FIG. 10D(2) is a reconstructed high-resolution image with digital refocusing and FIG. 10D(3) is a reconstructed high-resolution image without digital refocusing, for comparison.

Digital refocusing with FPI methods represents another significant improvement over conventional microscopes. Digital refocusing can correct: 1) cases where specimens are not perfectly aligned over the entire field-of-view and 2) for chromatic aberration digitally (assigning different defocused distances for different colors).

FIGS. 10E(1), 10E(2), 10E(3), 10E(4), 10E(5), 10E(6), 10E(7), and 10E(8) provide exemplary results from using an FPI system 10 to perform an FPI method with digital refocusing, according to an embodiment of the invention, that corrects for chromatic aberration in a blood smear and in a pathology slide. The FPI system 10 includes the FPI device 100(a) of FIG. 2B. FIG. 10E(1) is a representative low-resolution image of a blood smear as acquired by the FPI system 10. FIG. 10E(2) is a representative low-resolution image of a pathology slide as acquired by the FPI system 10. FIG. 10E(3) is a high-resolution image of the blood smear as computationally reconstructed with digital refocusing. FIG. 10E(4) is a high-resolution image of the pathology slide as computationally reconstructed with digital refocusing. FIG. 10E(5) is a high-resolution image of the blood smear as computationally reconstructed without digital refocusing. FIG. 10E(6) is a high-resolution image of the pathology slide as computationally reconstructed without digital refocusing. FIGS. 10E(7), and 10E(8) are conventional microscope images with a 20× objective lens of the blood smear and pathology slide respectively, for comparison.

C. FPI Methods with Digital Auto-Focusing

During operation of an FPI system 10, the z-position of the specimen 20 may not be known a priori. In certain embodiments, an FPI method may include a digital auto-focusing step that determines the z-position of the specimen 20 and uses this z-position to digitally refocus. For example, the FPI method of FIG. 6B may further comprise a step during or before step 1520 that computes the z-position of the specimen 20. The FPI system 10 may perform digital autofocusing by using the processor 210 to perform steps 1520 and 1540 in FIG. 6B using the computed z-position of the specimen. To compute the z-position of the specimen 20, the FPI method determines an auto-focusing index parameter. The auto-focusing index is defined by the following equation:Auto-focusing index: 1/Σabs(√{square root over (I_lf)}−√{square root over (I_lfm)})  (Eqn. 2)

Where: √{square root over (I_lf)} is the amplitude image from the low-pass filtering, and

√{square root over (I_lfm)} is the actual low-resolution measurement

The summation in Eqn. 2 is for all oblique incidence angles. After the FPI method computes the estimated z-position of the specimen 20, the FPI method can digitally refocus to the estimated z-position. In some cases, the high-resolution image solution has been found to converge better when using an accurate z-position.

FIGS. 10F(1), 10F(2), 10F(3), 10F(4), 10F(5), 10F(6), and 10F(7) include experimental results from using an FPI system 10 to perform an FPI method with digital autofocusing, according to an embodiment of the invention. This FPI method computes an auto-focusing index for computing the z-position of the specimen automatically. In this first case, the specimen was placed at z₀=−150 μm. FIGS. 10F(1), 10F(2), and 10F(3) are reconstructed images for the specimen at z₀=−150 μm. In the second case, the specimen was placed at z₀=−50 μm of an FPI device. FIGS. 10F(4), 10F(5), and 10F(6) are reconstructed images for the specimen at z₀=−50 μm. Using the FPI method, the FPI system 10 computed the auto-focusing index for different z-positions based on Eqn. 2. FIG. 10F(7) is a graph of the computed auto-focus index values for the different z-positions. The maximum value of the auto-focusing index indicates the estimated z-position of the specimen. As shown, the estimated positions based on the maximum calculated auto-focusing indexes are close to the actual positions of the specimen. Using an estimated positions, the FPI method can autofocus propagating the image to the estimated position.

D. FPI Methods with Digital Wavefront Correction

Although FPI methods do not require phase information as input, FPI methods of certain embodiments accommodate phase during iterative reconstruction. In these embodiments, the depth of focus can be extended beyond that of the objective lens using a numerical strategy to compensate for aberrations in the pupil function. Examples of such aberration compensation strategies can be found in Gutzler, T., Hillman, T. R., Alexandrov, S. A. and Sampson, D. D., “Coherent aperture-synthesis, wide-field, high-resolution holographic microscopy of biological tissue,” Opt. Lett. 35, pp. 1136-1138 (2010), and Colomb, T. et al., “Automatic procedure for aberration compensation in digital holographic microscopy and applications to specimen shape compensation,” Appl. Opt. 45, 851-863 (2006), which are hereby incorporated by reference in their entirety. The digital correction process digitally introduces a phase map to the coherent optical transfer function to compensate for aberrations at the pupil plane during the iterative image recovery process.

FIG. 9A is a flowchart of sub-steps of step 1500 of the FPI method of FIG. 6A, according to another embodiment of the invention. In this example, the FPI method includes digital wavefront correction. The FPI method incorporates digital wavefront compensation in the two multiplication steps 1605 and 1645. Specifically, step 1605 models the connection between the actual specimen profile and the captured intensity data (with includes aberrations) through multiplication with a pupil function: e^i·φ(kx,ky) by the processor 210. Step 1645 inverts such a connection to achieve an aberration-free reconstructed image. Specimen defocus is essentially equivalent to introducing a defocus phase factor to the pupil plane (i.e., a defocus aberration):e^i·φ(kx,ky)=eⁱ√{square root over (^{(2π/λ)2−kx2−ky2})}^·z0,k_x²+k_y²<(NA·2π/λ)²   (Eqn. 4)where k_x and k_y are the wavenumbers at the pupil plane, z₀ is the defocus distance, and NA is the numerical aperture of the optical element.

At step 1605, the processor 210 multiplies by a phase factor e^i·φ(kx,ky) in Fourier domain.

At step 1610, the processor 210 performs low-pass filtering of the high-resolution image √{square root over (I_h)}e^iφh in the Fourier domain to generate a low-resolution image √{square root over (I_l)}e^iφl for a particular plane wave incidence angle (θ_xⁱ, θ_yⁱ) with a wave vector (k_xⁱ, k_yⁱ). The Fourier transform of the high-resolution image is Ĩ_h and the Fourier transform of the low-resolution image for a particular plane wave incidence angle is Ĩ_l. In the Fourier domain, the FPI method filters a low-pass region from the spectrum Ĩ_h of the high-resolution image √{square root over (I_h)}e^iφh. In cases with an optical element in the form of an objective lens, this region is a circular aperture with a radius of NA*k₀, where k₀ equals 2π/λ (the wave number in vacuum), given by the coherent transfer function of an objective lens. In Fourier space, the location of the region corresponds to the incidence angle. For an oblique plane wave incidence with a wave vector (k_xⁱ, k_yⁱ), the region is centered about a position (−k_xⁱ,−k_yⁱ) in the Fourier domain of √{square root over (I_h)}e^iφh.

At step 1630, using the processor 210, the computed amplitude component √{square root over (I_lf)} of the low-resolution image at the in-focus plane, √{square root over (I_lf)}e^iφlf, is replaced with the square root of the low-resolution intensity measurement √{square root over (I_lfm)} measured by the radiation detector of the FPI device. This forms an updated low resolution target: √{square root over (I_lfm)}e^iφlf.

At step 1645, the processor 210 multiplies by an inverse phase factor e^{−i·φ(kx,ky)} in Fourier domain.

At step 1650, using the processor 210, a Fourier transform is applied to the updated target image propagated to the sample plane: √{square root over (I_ls)}e^iφls, and this data is updated in the corresponding region of high-resolution solution √{square root over (I_h)}e^iφh in the Fourier space corresponding to the corresponding to the incidence wave vector (k_xⁱ, k_yⁱ).

At step 1660, the processor 210 determines whether steps 1605 through 1650 have been completed for all incidence angles. If steps 1605 through 1650 have not been completed for all incidence angles, steps 1605 through 1650 are repeated for the next incidence angle.

In most embodiments, the neighboring regions in Fourier space, which are iteratively updated for each incidence angle, overlap each other. In the overlapping area between updated overlapping regions, the FPI system 10 has multiple samplings over the same Fourier space. The incidence angles determine the area of the overlapping area. In one embodiment, the overlapping area between neighboring regions may have an area that is between 2% to 99.5% of the area of one of the neighboring regions. In another embodiment, the overlapping area between neighboring regions may have an area that is between 65% to 75% of the area of one of the neighboring regions. In another embodiment, the overlapping area between neighboring regions may have an area that is about 65% of the area of one of the neighboring regions. In certain embodiments, each overlapping region has the same area.

At step 1670, the processor 210 determines whether the high-resolution solution has converged. For example, the processor 210 may determine whether the high-resolution solution may have converged to a self-consistent solution. In one case, the processor 210 compares the previous high-resolution solution of the previous iteration or initial guess to the present high-resolution solution, and if the difference is less than a certain value, the solution may have converged to a self-consistent solution. If processor 210 determines that the solution has not converged, then steps 1605 through 1670 are repeated. In one embodiment, steps 1605 through 1670 are repeated once. In other embodiments, steps 1605 through 1670 are repeated twice or more. If the solution has converged, the processor 210 transforms the converged solution in Fourier space to the spatial domain to recover a high-resolution image √{square root over (I_h)}e^iφh. If the processor 210 determines that the solution has converged at step 1570, then the process may proceed to optional step 1600.

FIG. 9B is a schematic diagram of an FPI device 100 implementing an FPI method with digital wavefront correction, according to an embodiment of the invention. As shown, the digital pupil function is introduced at steps 1605 and 1645 to model the connection between the actual specimen profile and the captured intensity data, which may exhibit aberrations caused by defocus. The FPI method with digital wavefront correction can also be used to correct for the spatially varying aberrations of the optical element (e.g., objective lens).

If the defocus distance is unknown, the FPI method can digitally adjust the ‘z’ parameter to different values based on a computation of the auto-focusing index from Eqn. 4. The FPI method can then reconstruct the corresponding images, and select the sharpest image. This approach can also be extended to image a tiled sample. In this case, the FPI method can digitally adjust the ‘z’ parameter to achieve acuity for each tiled region of the whole image and combine the in-focus regions to form a fully focused image of the full field of view.

In other embodiments, alternative digital multiplicative phase factors can be included in multiplication steps 1605 and 1645 to correct for a variety of aberrations, as long as the factors correctly model the employed optics.

E. Estimated Computational Cost of the Recovery Process

The computational cost of the recovery process of the illustrated example of FIG. 6B can be approximated by the computational cost of steps 1510, 1550, and 1570, assuming that the low-resolution images captured by the radiation detector contains n pixels and that N different light elements are used for illuminations. In step 1510 of the recovery process, a two-dimensional fast Fourier transform (FFT) is performed to generate the low-resolution image √{square root over (I_l)}e^iφl and the corresponding computational cost is n²·log(n). In step 1550, another FFT is performed to update the corresponding region of the Fourier space of √{square root over (I_hr)}e^iφhr and the corresponding computational cost is n²·log(n). At step 1560, the above computation is repeated once for all incidence angles, and thus the computational cost is N·2·n²·log(n). If step 1570 is repeated once, then the computational cost is 2·N·2·n²·log(n). Other steps in the recovery process are negligible compared to the above value. In summary, the computational complexity of the FPI method of FIG. 6B is 4·N·n²·log(n) where N is the total number of oblique incidences and n is the total number of pixels of the low-resolution image.

IV. Exemplary FPI systems with Color Imaging and/or Phase Imaging

A. Color Imaging

Color imaging capability is pivotal in pathology and histology. In certain embodiments, an FPI system 10 capable of color imaging comprises a variable illuminator with an LED having light elements that can provide red, green, and blue illuminations. The FPI method combines the high-resolution image results from red, green, and blue LED illumination into each corresponding color channel to form a final high-resolution color image. Three images are generated corresponding to red, green, and blue, which are combined to form a high resolution color image.

To demonstrate an implementation of this FPI system 10 for digital pathology applications, this color imaging system was used to acquire color high-resolution images of a pathology slide (human adenocarcinoma of breast section, Carolina). FIG. 11A(1) is a photograph illustrating the fields-of-view of a pathology slide for both 2× and 20× objective lenses of conventional microscopes, for comparison. For the case of the 2× objective lens, the FOV in the conventional microscope is ˜13.25 mm in diameter and the NA is 0.08, as shown by the images FIGS. 11A(2) and 11A(3). On the other hand, for the 20× objective lens, the NA of the 20× objective is 0.4, which is much higher than that of the 2× lens, while the FOV is only 1.1 mm in diameter, as shown in FIGS. 11A(4) and 11A(5). Using the color imaging FPI system 10 with a 2× objective lens of embodiments, the field-of-view is the same as the 2× case (i.e., 13.25 mm in diameter) while the maximum NA is −0.5, results in more than 1.8 gigapixel across the entire image by Nyquist rate. FIG. 11A(6) and FIG. 11A(7) are color images showing the field-of-view and corresponding maximum NA of an FPI system 10, according to an embodiment of the invention.

FIG. 11B includes results of using the color imaging FPI system 10, according to an embodiment of the invention. The row of figures corresponding to FIGS. 11B(1)-(18) are associated with the reconstructed images for red, green, and blue illumination respectively. FIG. 11B(19) is the raw color image. FIG. 11B(20) is the computationally reconstructed color image with digital refocusing by the FPI system 10. FIG. 11B(21) is an image captured by a 40× objective lens and a color image sensor for comparison. FIGS. 11C(1), 11C(2), 11C(3), 11C(4), 11C(5) and 11C(6) are color images showing a comparison of image quality between the color imaging FPI system 10 and different objective lenses. FIG. 11C(6) is an image computationally reconstructed by the FPI system 10. FIGS. 11C(1), 11C(2), 11C(3), 11C(4), and 11C(5) are images captured by a 2×, 4×, 10×, 20×, and 40× objective lenses, respectively, for comparison.

B. Phase Imaging

Phase imaging is useful in numerous applications. For example, phase profile of a histopathology slide may contain information about the molecular scale organization of tissue and can be used as an intrinsic marker for cancer diagnosis as discussed in Wang, Z., Tangella, K., Balla, A. and Popescu, G., “Tissue refractive index as marker of disease,” Journal of Biomedical Optics 16, 116017-116017 (2011), which is hereby incorporated by reference in its entirety. As another example, phase information of a sample can also be used to measure the optical path length of cells or organelles to determine the growth pattern of cell cultures or to perform blood screening, as discussed in Lue, N. et al., “Live Cell Refractometry Using Hilbert Phase Microscopy and Confocal Reflectance Microscopy,” The Journal of Physical Chemistry A, 113, pp. 13327-13330 (2009), Mir, M. et al., “Optical measurement of cycle-dependent cell growth,” Proceedings of the National Academy of Sciences 108, pp. 13124-13129 (2011), and Mir, M. et al., “Blood screening using diffraction phase cytometry,” Journal of Biomedical Optics 15, pp. 027016-027014 (2010), which are hereby incorporated by reference in their entirety.

Most conventional full-field phase imaging techniques use interferometry, which requires sophisticated and well-designed optical alignments. Compared to these conventional techniques, the FPI methods of embodiments are easy and cost-effective solution for researchers and clinicians to incorporate phase imaging functionality into their current microscope systems. The Figures in FIG. 11D includes phase and color images from using an FPI method with the color imaging FPI system 10 for both a pathology slide and blood smear, according to an embodiment of the invention. FIGS. 11D(1) and 11D(2) are the low-resolution images. FIGS. 11D(3) and 11D(4) are the high-resolution intensity images. FIGS. 11D(5) and 11D(6) are the high-resolution phase images. FIGS. 11D(7) and 11D(8) are the high-resolution color images. FIGS. 11D(9) and 11D(10) are images from a conventional microscope with a 20× objective lens for comparison. The phase image is generated by taking the phase angle of the complex high-resolution image.

V. Further Simulation Results

Some results of numerical simulations of FPI systems performing FPI methods, according to embodiments of the invention, are provided in this section and elsewhere in the disclosure. In many embodiments, the FPI systems use FPI methods to reconstruct a high-resolution image with the use of a low NA lens.

As an example, FIG. 5A illustrates results from a numerical simulation of an FPI system using an FPI method, according to an embodiment of the invention. FIG. 5A includes two images (left-hand side) that represent the simulated specimen intensity profile I(x,y) and phase profile φ(x,y). The pixel size of these two input profiles is 275 nm and the wavelength of the simulated incidence is 632 nm. In this simulation, the method was performed using an FPI device 100(a) having a plurality of stationary light elements in the form of a two-dimensional array of light elements. During the measurement process, the specimen was illuminated by plane waves at 137 different incident angles, and filtered by a 2× objective lens (0.08 NA), and then captured by an image sensor with 5.5 μm pixel size. In this simulation, the sample is assumed to be placed at the in-focus position of the objective lens (i.e. at the in-focus plane 122). The resulting low-resolution images, added with 1% random speckle noise, are shown in the middle part of FIG. 5A. Based on these low-resolution images, a recovery process was used to computationally reconstruct a high-resolution image with a maximum NA of 0.5 in Fourier space. In the recovery process, the intensity of the low-pass filtered image is replaced by the actual low-resolution measurement and the corresponding Fourier space of the high-resolution reconstruction is updated accordingly. Intensity and phase profiles of the reconstructed high-resolution image are shown in the right-hand side of FIG. 5A. The high-resolution images of the specimen may be recovered without involving phase measurement in the data acquisition measurement process.

FIGS. 5B(1), 5B(2), 5B(3), 5B(4), 5B(5), 5B(6), 5B(7), 5B(8), 5B(9), 5B(10), 5B(11), and 5B(12) illustrate more results from the numerical simulation of the FPI system using an FPI method, according to the embodiment discussed with respect to FIG. 5A. FIGS. 5B(1), 5B(2), 5B(3), 5B(4), 5B(5), 5B(6), 5B(7), 5B(8), and 5B(9) are images of 9 low-resolution measurements out of the 137 taken during this simulation. FIGS. 5B(10) and 5B(11) are computationally reconstructed high-resolution intensity and phase images respectively. FIG. 5B(12) is the Fourier space of the recovered image. The corresponding regions of the low-resolution images are highlighted.

VI. Throughput Enhancement Factor

Based on the pixel size ratio between the largest pixel sizes of the raw low-resolution images and the largest pixel sizes of the high-resolution images, an enhancement factor of using an FPI system can be expressed as:Enhancement factor=2·NA_syn/NA_obj  (Eqn. 3)The larger the Enhancement factor, the higher the system throughput.

In certain embodiments, the FPI system 10 may have a certain sampling condition to provide a particular enhancement factor. Generally, the sampling condition may be associated with two questions: 1) given the NA of an objective lens, what is the largest pixel size that can be used for acquiring the low resolution intensity images; 2) given the synthetic NA of the reconstructed image, what is the largest pixel size that can be used for representing the reconstructed intensity image. Since the FPI method of embodiments can recover both intensity and phase information, the answer to question 1 is the same as the sampling condition for coherent optical systems: λ/(2·NA_obj). For question 2, the synthetic NA is for the electric field E (with amplitude and phase). The final reconstruction, on the other hand, is for the intensity profile I (I=E·E*, where ‘*’ denotes for complex conjugate). Such a multiplication of electric field in the spatial domain corresponds to a convolution operation in the Fourier domain. As such, the passband of the reconstructed intensity image double in the Fourier domain. Therefore, the largest pixel size that can be used for representing the reconstructed intensity image is λ/(4·NA_obj) at the sample plane.

In certain implementations, an FPI device 100(a) includes an optical element in the form of a 2× objective with 0.08 NA and radiation detector in the form of an image sensor with 5.5 μm pixel size. From the answer to question 2, the largest pixel size that can be used at the image plane is 2·λ/(2·NA_obj)=5.88 μm for blue light. The 5.5 μm pixel size of the image sensor is, therefore, in accord with such a sampling condition. On the other hand, based on the answer to question 2, pixel sizes (at the sample plane) of the reconstructed image are 0.34 μm, 0.30 μm and 0.28 μm for red, green and blue wavelengths. For simplicity, a reconstructed pixel size of 0.275 μm was used for these three wavelengths in the implementation, corresponding to an enhancement factor of 10.

VII. Advantages of the FPI system

Embodiments of the invention may provide one or more of the following technical advantages. Some examples of these advantages are provided below.

1. Object Support in Fourier Domain.

In one aspect, FPI methods impose object support constraints in the Fourier domain, which provides large FOV and higher signal-to-noise ratio (with focusing elements) without mechanical scanning. In conventional ptychography systems, object support is provided by the confined illumination in the spatial domain. In these conventional systems, the specimen must be mechanically scanned through the desired field-of-view.

2. No phase measurements

In most embodiments, the FPI methods do not require phase measurements of the specimen. Conventional interfermometric synthetic aperture microscopes require phase measurements in their detection schemes, as can be described in Rodenburg, J. M. and Bates, R. H. T. “The theory of super-resolution electron microscopy via Wigner-distribution deconvolution,” Phil. Trans. R. Soc. Lond. A 339, 521-553 (1992), H. M. L. and Rodenburg, J. M., “Movable aperture lensless transmission microscopy, a novel phase retrieval algorithm,” Phys. Rev. Lett. 93, 023903 (2004), Rodenburg, J. M. et al., “Hard-X-ray lensless imaging of extended objects,” Phys. Rev. Lett. 98, 034801 (2007), Thibault, P. et al., “High-resolution scanning X-ray diffraction microscopy,” Science 321, 379-382 (2008), Dierolf, M. et al., “Ptychographic coherent diffractive imaging of weakly scattering specimens,” New J. Phys. 12, 035017 (2010), Maiden, A. M., Rodenburg, J. M. and Humphry, M. J., “Optical ptychography: a practical implementation with useful resolution,” Opt. Lett. 35, 2585-2587 (2010), Humphry, M., Kraus, B., Hurst, A., Maiden, A. and Rodenburg, J., “Ptychographic electron microscopy using high-angle dark-field scattering for sub-nanometre resolution imaging,” Nat. Commun. 3, 730 (2012), which are hereby incorporated by reference in their entirety. Because no measured phase information is needed, The FPI system eliminates the design challenges associated with conventional interferometric detection schemes.

3. Spatial-bandwidth product beyond the physical limitations of optical element

One advantage may be that the FPI systems of certain embodiments may increase the spatial-bandwidth product beyond the physical limitations of its optics while avoiding the drawbacks of previous attempts to do so with conventional devices, in a simple and cost-effective manner. These systems use FPI methods to decouple field-of-view from resolution of the optics, which allows these systems to generate a high-resolution image with a low NA lens. Thus, the throughput of these FPI systems is not limited by the spatial-bandwidth product of their optics. In one exemplary embodiment, an FPI system has been implemented that produces a 1.6 gigapixel system with a maximum NA of 0.5, a field-of-view of 120 mm², and a resolution-invariant imaging depth of 0.3 mm.

4. Scalable Throughput.

As discussed, the resolution and FOV are not coupled in FPI systems of embodiments. In one embodiment, an FPI system can be scaled for a desired throughput by using an objective lens having a particular NA and by using a particular number N of illuminations. For example, the throughput of an FPI system may be increased by reducing the NA of the objective lens and/or by increasing the number N of illuminations. In one embodiment example, an FPI system provides two orders of magnitude higher throughput than existing bright-field microscopes without the use of mechanical scanning and/or phase measurements

5. Modular form that can be Readily Implemented With Conventional Microscope

In certain cases, components of the FPI system may be in modular form to be readily implemented with components of a conventional microscope or other conventional imaging device. These implementations may have the potential to broadly impact digital pathology, hematology, phytotomy, immunochemistry and forensic photography. For example, modular components of an FPI system may include a variable illuminator (e.g., a simple light-emitting diode (LED) matrix) and a processor with instructions for executing the FPI method.

6. No mechanical scanning.

In certain embodiments, FPI methods do not need mechanical scanning. Unlike conventional synthetic-aperture and scanning-based wide FOV microscopy techniques, mechanical scanning is not required for FPI methods of embodiments. As such, it simplifies the platform design, reduces associated costs and allows for a higher throughput limit.

7. Capable of color imaging.

Another advantage may be that the FPI system of certain embodiments can be used to generate color images, which is pivotal in pathology and histology applications. In embodiments, an FPI system may be capable of color imaging by having a variable illuminator that provides illumination of different colors (e.g., color LED matrix). If digital refocusing is implemented, the computationally reconstructed image can remain free of any chromatic aberrations of the optics.

8. Digital Refocusing and Expanded Depth of Focus

To achieve optimal resolution in a conventional microscope, a stage must be used to mechanically bring an out-of-focus specimen into focus. In certain embodiments, FPI methods can computationally refocus the image digitally, rather than mechanically. This aspect is especially useful for cases where specimens are not perfectly aligned over the entire FOV. Using digital refocusing, the FPI system can expand the depth of focus beyond the optical limitations of the optical element. For example, an FPI method of one embodiment can been used to extend the depth of focus of the objective lens of a conventional microscope from about 80 μm to about 0.3 mm. This increased depth of focus can provide the additional advantage of providing a large tolerance to microscope slide placement errors, which improves accuracy over conventional systems.

9. Phase Imaging Capability.

A phase profile of a biological sample contains information about the molecular scale organization of sample and can be used as an intrinsic marker for different applications. The ability to perform phase imaging is also useful for digital image processing, such as cell segmentation and cell counting. Conventional full-field phase imaging techniques require fairly sophisticated and well-designed optical alignments. In certain embodiments, an FPI method provides an easy and cost-effective solution for researchers and clinicians to incorporate phase imaging functionality into their current microscope systems. The FPI method can reconstruct the complex image of the sample. In certain embodiments, the FPI method can generate a phase image by taking the phase angle of the complex image.

10. X-Ray and THz applications

In certain embodiments, FPI methods can be extended to Terahertz and X-ray imaging applications, where lenses are poor and of very limited numerical aperture.

VIII. FPI Systems for X-Ray Imaging

In certain embodiments, an FPI system 10 may be configured for X-ray imaging. In these embodiments, the Fourier ptychographic X-ray imaging system 10 comprises an FPI device with an X-ray radiation source (e.g., X-ray tube and metal target).

In one embodiment, a Fourier ptychographic X-ray imaging system 10 may include the FPI device 100(c) shown in FIG. 4A, with components configured for use with X-ray radiation. In this embodiment, the FPI device 110(c) comprises an assembly 170 that may be mounted as a rigid assembly onto a stage (e.g., a goniometer stage) for translating and/or rotating the assembly 170 with respect to a stationary light element 112 directing X-ray radiation to provide X-ray radiation to a specimen 20 from a plurality of N incidence angles. The FPI device 100(c) comprises a mechanism 160 for translating and/or rotating the assembly 170 with respect to the light element 112. The mechanism 170 may be affixed to a stationary base. The FPI device 100(c) also comprises an optical element 130(c) and a radiation detector 140(c). Both of these components are designed for X-ray radiation i.e. an X-ray optical element and an X-ray radiation detector respectively. The X-ray optical element may be, for example, a micro zone plate or a grazing incidence mirror. A micro zone plate can project a full-field image to the X-ray radiation detector. The X-ray radiation detector may be, for example, an X-ray sensitive CCD. An example of a micro zone plate and an X-ray sensitive CCD can be found in Chao, W., Harteneck, B., Liddle, A., Anderson, E., and Attwood, D., “Soft X-ray microscopy at a spatial resolution better than 15 nm,” Nature, vol. 435 (Jun. 30, 2005), which is hereby incorporated by reference in its entirety. In this embodiment, the light element 112 may comprise a stationary X-ray radiation source.

In another embodiment, a Fourier ptychographic X-ray imaging system 10 may include the FPI device 100(d) shown in FIG. 4B, with components configured for use with X-ray radiation. In this embodiment, the FPI device 110(d) comprises a light element 112 directing X-ray radiation to the specimen 20 from a plurality of N incidence angles. The FPI device 100(d) rotates the light element 112 relative to the specimen. For example, the light element 112 may pivot about one or more axes. The FPI device 100(d) also comprises an optical element 130(b) and a radiation detector 140(b). In this embodiment, the optical element 130(b) and radiation detector 140(b) may be designed for X-ray radiation i.e. may be an X-ray optical element (e.g. micro zone plate or grazing incidence mirror) and an X-ray radiation detector (e.g., X-ray sensitive CCD) respectively. In this embodiment, the light element 112 may comprise an X-ray radiation source.

The Fourier ptychographic X-ray imaging system 10 may perform the FPI method described with respect to FIG. 6A. At step 1100, the FPI method provides illumination to the specimen area from a plurality of N incidence angles by translating and/or rotating the assembly 170 with respect to a stationary light element 112 directing X-ray radiation to provide X-ray radiation to a specimen 20 from a plurality of N incidence angles.

By implementing X-ray imaging, these Fourier ptychographic X-ray imaging systems 10 provide additional capabilities. For example, these systems may provide larger penetration depths into the specimen. As another example, X-ray imaging may allow for elemental and chemical identification on a scale of 10 nm or less.

IX. Subsystems

FIG. 12 is a block diagram of subsystems that may be present in a FPI system 10, according to embodiments. For example, the FPI system 10 includes a processor 210. The processor 210. The processor 210 may be a component of the FPI device 100 in some cases. The processor 210 may be a component of the radiation detector 140 in some cases.

The various components previously described in the Figures may operate using one or more of the subsystems to facilitate the functions described herein. Any of the components in the Figures may use any suitable number of subsystems to facilitate the functions described herein. Examples of such subsystems and/or components are shown in a FIG. 12. The subsystems shown in FIG. 12 are interconnected via a system bus 2425. Additional subsystems such as a printer 2430, keyboard 2432, fixed disk 2434 (or other memory comprising computer readable media), display 230, which is coupled to display adapter 2438, and others are shown. Peripherals and input/output (I/O) devices, which couple to I/O controller 2440, can be connected by any number of means known in the art, such as serial port 2442. For example, serial port 2442 or external interface 2444 can be used to connect the computing device 200 to a wide area network such as the Internet, a mouse input device, or a scanner. The interconnection via system bus 2425 allows the processor 210 to communicate with each subsystem and to control the execution of instructions from system memory 2446 or the fixed disk 2434, as well as the exchange of information between subsystems. The system memory 2446 and/or the fixed disk 2434 may embody the CRM 220 in some cases. Any of these elements may be present in the previously described features.

In some embodiments, an output device such as the printer 2430 or display 230 of the FPI system 10 can output various forms of data. For example, the FPI system 10 can output 2D color/monochromatic images (intensity and/or phase), data associated with these images, or other data associated with analyses performed by the FPI system 10.

Modifications, additions, or omissions may be made to any of the above-described FPI methods and their associated features, e.g., features described with respect to FIGS. 6A and 6B and other illustrations, without departing from the scope of the disclosure. Any of the FPI methods described above may include more, fewer, or other features without departing from the scope of the disclosure. Additionally, the steps of the described features may be performed in any suitable order without departing from the scope of the disclosure.

It should be understood that the present invention as described above can be implemented in the form of control logic using computer software in a modular or integrated manner. Based on the disclosure and teachings provided herein, a person of ordinary skill in the art will know and appreciate other ways and/or methods to implement the present invention using hardware and a combination of hardware and software.

Any of the software components or functions described in this application, may be implemented as software code to be executed by a processor using any suitable computer language such as, for example, Java, C++ or Perl using, for example, conventional or object-oriented techniques. The software code may be stored as a series of instructions, or commands on a CRM, such as a random access memory (RAM), a read only memory (ROM), a magnetic medium such as a hard-drive or a floppy disk, or an optical medium such as a CD-ROM. Any such CRM may reside on or within a single computational apparatus, and may be present on or within different computational apparatuses within a system or network.

Although the foregoing disclosed embodiments have been described in some detail to facilitate understanding, the described embodiments are to be considered illustrative and not limiting. It will be apparent to one of ordinary skill in the art that certain changes and modifications can be practiced within the scope of the appended claims.

One or more features from any embodiment may be combined with one or more features of any other embodiment without departing from the scope of the disclosure. Further, modifications, additions, or omissions may be made to any embodiment without departing from the scope of the disclosure. The components of any embodiment may be integrated or separated according to particular needs without departing from the scope of the disclosure.

Claims

1. A Fourier ptychographic imaging device, comprising: an array of light elements, each light element comprising at least one light source, wherein the array of light elements is configured to provide illumination sequentially from each of a plurality of incidence angles to a specimen being imaged during operation;an optical element comprising at least one lens, wherein the optical element is configured to receive light issuing from the specimen being illuminated and is configured to filter the light received by propagating light based on an acceptance angle of the at least one lens;a radiation detector configured to receive light from the optical element and to acquire a plurality of variably-illuminated, low-resolution intensity images of the specimen, the radiation detector configured to acquire each low-resolution intensity image based on light issuing directly from the specimen and filtered by the optical element while the specimen is illuminated at one of the plurality of incidence angles; anda processor configured to reconstruct a high-resolution image of the specimen by iteratively determining a high-resolution solution, wherein reconstructing the high-resolution image comprises computationally recovering phase image information from the variably-illuminated, low-resolution intensity images, wherein iteratively determining the high-resolution solution comprises:(i) using a current high-resolution solution to generate data in each a plurality of regions in the Fourier domain, the plurality of regions associated with the plurality of incidence angles;(ii) inverse Fourier transforming the data in a region of the plurality of regions to generate image data, replacing amplitudes of the image data with square root of an intensity measurement from at least one of the plurality of variably-illuminated, low-resolution intensity images corresponding to a location of the region in Fourier space to produce updated image data, Fourier transforming the updated image data to produce Fourier-transformed updated data, and updating data in the region of the Fourier domain with the Fourier-transformed updated data;(iii) repeating operation in (ii) for each region of the plurality of regions in the Fourier domain to determine an updated high-resolution solution; and(iv) repeating operations in (iii) if a comparison of data derived from the updated high-resolution solution and data derived from the current high-resolution solution is more than a value,wherein the high-resolution image reconstructed has higher resolution that the low-resolution intensity images.

2. The Fourier ptychographic imaging device of claim 1, wherein the optical element is an objective that comprises the at least one lens.

3. The Fourier ptychographic imaging device of claim 1, wherein one or more sets of adjacent regions of the plurality of regions in the Fourier domain are overlapping in Fourier space.

4. The Fourier ptychographic imaging device of claim 1, wherein a difference between the two most adjacent incidence angles in the plurality of incidence angles is between 10% and 90% of the acceptance angle corresponding to a numerical aperture of the at least one lens.

5. The Fourier ptychographic imaging device of claim 1, wherein a difference between the two most adjacent incidence angles in the plurality of incidence angles is between 33% and 66% the acceptance angle corresponding to a numerical aperture of the at least one lens.

6. The Fourier ptychographic imaging device of claim 1, wherein a difference between the two most adjacent incidence angles in the plurality of incidence angles is less than 76% of the acceptance angle corresponding to a numerical aperture of the at least one lens.

7. The Fourier ptychographic imaging device of claim 1, wherein the array of light elements is a two-dimensional matrix of light elements, each light element providing illumination from one of the plurality of incidence angles.

8. The Fourier ptychographic imaging device of claim 7, wherein each light element is a set of one or more light-emitting diodes.

9. The Fourier ptychographic imaging device of claim 1, wherein the array of light elements is a hexagonal array of light elements, each light element providing illumination from one of the plurality of incidence angles.

10. The Fourier ptychographic imaging device of claim 9, wherein each light element is a set of one or more light-emitting diodes.

11. The Fourier ptychographic imaging device of claim 3, wherein the one or more sets of adjacent regions overlap by between 20% and 90% in area.

12. The Fourier ptychographic imaging device of claim 3, wherein the one or more sets of adjacent regions overlap by between 2% and 99.5% in area.

13. The Fourier ptychographic imaging device of claim 3, wherein the one or more sets of adjacent regions overlap by between 33% and 66% in area.

14. The Fourier ptychographic imaging device of claim 1, wherein the processor is also configured to automatically refocus the high-resolution image to an in-focus plane of the specimen.

15. The Fourier ptychographic imaging device of claim 1, wherein the processor is part of the radiation detector.

16. The Fourier ptychographic imaging device of claim 1, further comprising a display for displaying the high-resolution image.

17. A method of Fourier ptychographic imaging, comprising: causing a specimen to be illuminated from each of a plurality of incidence angles using an array of light elements, each light element comprising at least one light source, wherein light issuing from the specimen is filtered during illumination using an optical element comprising at least one lens, wherein the light is filtered based on an acceptance angle of the at least one lens;causing a plurality of variably-illuminated, low-resolution intensity images of the specimen to be acquired using a radiation detector that receives light from the optical element, wherein each low-resolution intensity image is captured based on light issuing from the specimen and filtered by the optical element while the specimen is illuminated at one of the plurality of incidence angles; andreconstructing a high-resolution image of the specimen by iteratively determining a high-resolution solution, wherein reconstructing the high-resolution image comprises computationally recovering phase image information from the variably-illuminated, low-resolution intensity images, wherein the high-resolution image reconstructed has higher resolution that the low-resolution intensity images, wherein iteratively determining the high-resolution solution comprises:(i) using a current high-resolution solution to generate data in each of a plurality of regions in the Fourier domain, the plurality or regions associated with the plurality of incidence angles;(ii) inverse Fourier transforming the data in a region of the plurality of regions to generate image data, replacing amplitudes of the image data with square root of an intensity measurement from at least one of the plurality of variably-illuminated, low-resolution intensity images associated with an illumination angle corresponding to a location of the region in Fourier space to produce updated image data, Fourier transforming the updated image data to produce Fourier-transformed updated data, and updating data in the region of the Fourier domain with the Fourier-transformed updated imago data;(iii) repeating operation in (ii) for each region of the plurality of regions in the Fourier domain to determine an updated high-resolution solution; and(iv) repeating operations in (iii) if a comparison of data derived from the updated high-resolution solution and data derived from the current high-resolution solution is more than a value.

18. The method of Fourier ptychographic imaging of claim 17, wherein one or more sets of adjacent regions of the plurality of regions in the Fourier domain are overlapping regions in Fourier space.

19. The method of Fourier ptychographic imaging of claim 18, wherein reconstructing the high-resolution image of the specimen comprises: dividing each variably-illuminated, low-resolution intensity image into a plurality of variably-illuminated, low-resolution intensity tile images;reconstructing a high-resolution tile image for each tile; andcombining the high-resolution tile images of the tiles.

20. The method of Fourier ptychographic imaging of claim 17, further comprising refocusing the high-resolution image.

21. The method of Fourier ptychographic imaging of claim 17, further comprising automatically focusing the high-resolution image.

22. The method of Fourier ptychographic imaging of claim 17, further comprising inverse Fourier transforming data of the updated high-resolution solution to produce the high-resolution image of the specimen.

23. The method of Fourier ptychographic imaging of claim 17, wherein (ii) further comprises multiplying the data in each of the plurality of regions with a phase factor.

24. A method of Fourier ptychographic imaging, comprising: receiving a plurality of variably-illuminated, low-resolution intensity images of a specimen from an image sensor, wherein each low-resolution intensity image is captured based on light issuing from the specimen and filtered by an optical element while the specimen is illuminated sequentially at each respective one of a plurality of incidence angles; andreconstructing a high-resolution image of the specimen by iteratively determining a high-resolution solution, wherein reconstructing the high-resolution image comprises computationally recovering phase image information from the variably-illuminated, low-resolution intensity images, wherein the high-resolution image reconstructed has higher resolution that the low-resolution intensity images, wherein iteratively determining the high-resolution solution comprises: (i) using a current high-resolution solution to generate data in each of a plurality of regions in the Fourier domain, the plurality of regions associated with the plurality of incidence angles;(ii) inverse Fourier transforming the data a region of the plurality of regions to generate image data, replacing amplitudes of the image data with square root an intensity measurement from at least one of the plurality of variably-illuminated, low-resolution intensity images associated with an illumination angle corresponding to a location of the region in Fourier space to produce updated image data, Fourier transforming the updated image data to produce Fourier-transformed updated data, and updating data the region of the Fourier domain with the Fourier-transformed updated data;(iii) repeating operation in (ii) for each region of the plurality of regions in the Fourier domain to determine an updated high-resolution solution; and(iv) repeating operations in (iii) if a comparison of data derived from the updated high-resolution solution and data derived from the current high-resolution solution is more than a value.

25. The method of Fourier ptychographic imaging of claim 24, one or more sets of adjacent regions of the plurality of regions in the Fourier domain are overlapping in Fourier space.

26. The method of Fourier ptychographic imaging of claim 25, wherein reconstructing the high-resolution image of the specimen comprises: dividing each variably-illuminated, low-resolution intensity image into a plurality of variably-illuminated, low-resolution intensity tile images;reconstructing a high-resolution tile image for each tile; andcombining the high-resolution tile images of the tiles.

27. The method of Fourier ptychographic imaging of claim 24, further comprising inverse Fourier transforming data of the updated high-resolution solution to produce the high-resolution image of the specimen.

28. The method of Fourier ptychographic imaging of claim 24, wherein (i) further comprises multiplying the data in each of the plurality of regions with a phase factor.

29. The method of Fourier ptychographic imaging of claim 25, wherein each region corresponds to an approximate optical transfer function of the optical element.

30. The Fourier ptychographic imaging device of claim 1, wherein (i) further comprises multiplying the data in each of the plurality of regions with a phase factor.

31. The Fourier ptychographic imaging device of claim 1, wherein iteratively determining the high-resolution solution further comprises inverse Fourier transforming data of the updated high-resolution solution to produce the high-resolution image of the specimen.