Luminal background cleaning

Abstract

Apparatus and methods are described for use with an input angiogram image of a device inserted inside a portion of subject's body, the angiogram image being acquired in the presence of contrast agent. At least one processor (11) includes background-image-generation functionality (13) configured to generate a background image in which a relative value is assigned to a first pixel with respect to a second pixel, at least partially based upon relative values of surroundings of the first pixel and the surroundings of the second pixel in the input image. Cleaned-image-generation functionality (14) generates a cleaned image in which visibility of the radiopaque portions of the device is increased relative to the input image, by dividing the input image by the background image. Output-generation functionality (15) drives a display (16) to display an output based upon the cleaned image. Other applications are also described.

Description

FIELD OF EMBODIMENTS OF THE INVENTION

Applications of the present invention generally relate to medical image processing. Specifically, applications of the present invention relate to background cleaning in images of body lumens and body cavities.

BACKGROUND OF THE INVENTION

Vascular catheterizations, such as coronary catheterizations, are frequently-performed medical interventions. Such interventions are typically performed in order to diagnose the blood vessels for potential disease, and/or to treat diseased blood vessels. Typically, in order to facilitate visualization of blood vessels, the catheterization is performed under extraluminal imaging. Typically, and in order to highlight the vasculature during such imaging, a contrast agent is periodically injected into the applicable vasculature. The contrast agent typically remains in the vasculature only momentarily. During the time that the contrast agent is present in the applicable vasculature, the contrast agent typically hides, in full or in part, or obscures, devices positioned or deployed within that vasculature.

The following articles do not necessarily pertain to medical procedures or body organs, but yet serve as a useful technical background.

An article entitled “Nonlocal linear image regularization and supervised segmentation,” by Gilboa and Osher (SIAM Multiscale Modeling & Simulation, volume 6, issue 2, pp. 595-630, 2007), which is incorporated herein by reference, describes how a nonlocal quadratic functional of weighted differences is examined. The weights are based on image features and represent the affinity between different pixels in the image. By prescribing different formulas for the weights, one can generalize many local and nonlocal linear de-noising algorithms, including the nonlocal means filter and the bilateral filter. In this framework one can show that continuous iterations of the generalized filter obey certain global characteristics and converge to a constant solution. The linear operator associated with the Euler-Lagrange equation of the functional is closely related to the graph Laplacian. Thus, the steepest descent for minimizing the functional as a nonlocal diffusion process may be determined. This formulation allows a convenient framework for nonlocal variational minimizations, including variational denoising, Bregman iterations and the recently-proposed inverse-scale-space. The authors demonstrate how the steepest descent flow can be used for segmentation. Following kernel based methods in machine learning, the generalized diffusion process is used to propagate sporadic initial user's information to the entire image. The process is not explicitly based on a curve length energy and thus can cope well with highly non-convex shapes and corners. Reasonable robustness to noise is achieved.

An article entitled “Nonlocal Operators with Applications to Image Processing,” by Gilboa and Osher (SIAM Multiscale Modeling & Simulation, volume 7, issue 3, pp. 1005-1028, 2008), which is incorporated herein by reference, describes the use of nonlocal operators to define types of flows and functionals for image processing and other applications. The authors describe a main advantage of the technique over classical Partial-Differential-Equation-based (PDE-based) algorithms as being the ability to handle better textures and repetitive structures. This topic can be viewed as an extension of spectral graph theory and the diffusion geometry framework to functional analysis and PDE-like evolutions. Some possible applications and numerical examples of the technique are provided, as is a general framework for approximating Hamilton-Jacobi equations on arbitrary grids in high dimensions, e.g., for control theory.

An article entitled “Non-local regularization of inverse problem,” by Peyre, Bougleux, and Cohenin (Lecture Notes in Computer Science, 2008, Volume 5304/2008, pp. 57-68), which is incorporated herein by reference, proposes a new framework to regularize linear inverse problems using the total variation on non-local graphs. A nonlocal graph allows adaptation of the penalization to the geometry of the underlying function to recover. A fast algorithm computes, iteratively, both the solution of the regularization process and the non-local graph adapted to this solution.

An article entitled “The split Bregman method for L1 regularized problems,” by Goldstein and Osher (SIAM Journal on Imaging Sciences, Volume 2, Issue 2, pp. 323-343), which is incorporated herein by reference, notes that the class of 11-regularized optimization problems has received much attention recently because of the introduction of “compressed sensing,” which allows images and signals to be reconstructed from small amounts of data. Despite this recent attention, many 11-regularized problems still remain difficult to solve, or require techniques that are very problem-specific. The authors show that Bregman iteration can be used to solve a wide variety of constrained optimization problems. Using this technique, the authors propose a “Split Bregman” method, which can solve a very broad class of 11-regularized problems.

In an article entitled “Bregmanized nonlocal regularization for reconvolution and sparse reconstruction,” by Zhang, Burgery, Bresson, and Osher (SIAM Journal on Imaging Sciences, Volume 3, Issue 3, July 2010), which is incorporated herein by reference, the authors propose two algorithms based on Bregman iteration and operator splitting technique for nonlocal TV regularization problems. The convergence of the algorithms is analyzed and applications to deconvolution and sparse reconstruction are presented.

SUMMARY OF EMBODIMENTS

Some applications of the present invention are applied to medical procedures performed, in whole or in part, on or within luminal body structures or body cavities. For some applications, apparatus and methods are provided for facilitating the visualization of devices positioned or deployed within a lumen or cavity at a time when the lumen or cavity is injected with contrast agent.

It should be appreciated that while using coronary catheterization as a primary example, applications of the current invention may be applied to any medical procedure in which a medical device is positioned or deployed within a body lumen or cavity, while the lumen or cavity is injected with a substance for the purpose of better discerning that lumen or cavity by means of medical imaging. Such lumens or cavities include, without limitation, any lumen or cavity of the cardiovascular system, the gastro-intestinal tract, the respiratory tract, the urinary tract, the nasal cavities, and/or any other bodily lumen or cavity.

There is therefore provided, in accordance with some applications of the present invention, apparatus for use with an input angiogram image of a device inserted inside a portion of a body of a subject, the device including radiopaque portions thereof, the angiogram image being acquired in the presence of contrast agent within the body portion, and for use with a display, the apparatus including:

at least one processor that includes:

• • background-image-generation functionality configured to generate a background image in which a relative value is assigned to a first pixel with respect to a second pixel, at least partially based upon relative values of surroundings of the first pixel and the surroundings of the second pixel in the input image;
  • cleaned-image-generation functionality configured to generate a cleaned image in which visibility of the radiopaque portions of the device is increased relative to the input image, by dividing the input image by the background image; and
  • output-generation functionality configured to drive the display to display an output based upon the cleaned image.

For some applications, the background-image-generation functionality is configured to generate the background image by:

in the background image, assigning pixel values to the first pixel and the second pixel that are more similar to one another, than the similarity of pixel values that are assigned to a third pixel and a fourth pixel,

based upon the first pixel and the second pixel having more similar surroundings to one another in the input image, than a similarity of surroundings of the third pixel and the fourth pixel to one another in the input image.

For some applications, the background-image-generation functionality is configured to generate the background image by assigning values to the first and second pixels based upon values of the first pixel and the second pixel in the input image.

For some applications, the output-generation functionality is configured to drive the display to display the cleaned image.

For some applications:

the input angiogram image of the device includes a plurality of input angiogram images of the device,

the cleaned-image-generation functionality is configured to generate a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images, and

the output-generation functionality is configured to:

• • generate a stabilized image stream by image tracking the cleaned images with respect to each other, based upon locations of the radiopaque portions of the device in the cleaned images, and
  • drive the display to display the stabilized image stream.

For some applications:

the input angiogram image of the device includes a plurality of input angiogram images of the device,

the cleaned-image-generation functionality is configured to generate a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images, and

the output-generation functionality is configured to:

• • generate a stabilized image stream by image tracking the input images with respect to each other, based upon locations of the radiopaque portions of the device in the corresponding cleaned images, and
  • drive the display to display the stabilized image stream.

For some applications:

the input angiogram image of the device includes a plurality of input angiogram images of the device,

the cleaned-image-generation functionality is configured to generate a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images, and

the output-generation functionality is configured to:

• • generate an enhanced image frame by:
    • aligning the cleaned images with each other, based upon locations of the radiopaque portions of the device in the cleaned images, and
    • generating an averaged image frame based upon the aligned images, and
  • drive the display to display the enhanced image frame.

For some applications:

the input angiogram image of the device includes a plurality of input angiogram images of the device,

the cleaned-image-generation functionality is configured to generate a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images, and

the output-generation functionality is configured to:

• • generate an enhanced image frame by:
    • aligning the input images with each other, based upon locations of the radiopaque portions of the device in the corresponding cleaned images, and
    • generating an averaged image frame based upon the aligned images, and
  • drive the display to display the enhanced image frame.

For some applications:

the portion of the subject's body includes a lumen of the subject's body,

the device includes an endoluminal data-acquisition device configured to acquire endoluminal data points while the device is at respective locations within the lumen, and

the output-generation functionality is configured:

• • based upon locations of the radiopaque portions of the device in the cleaned image, to determine that a given one of the endoluminal data points corresponds to a given location within the lumen, and
  • to drive the display to display an output, in response to the determining.

For some applications:

the portion of the subject's body includes a lumen of the subject's body,

the device includes an endoluminal data-acquisition device configured to acquire endoluminal data points while the device is at respective locations within the lumen, and

the output-generation functionality is configured:

based upon locations of the radiopaque portions of the device in the cleaned image, to determine that the endoluminal device is at a given location within the lumen, and

in response to the determining that the endoluminal device is at the given location within the lumen, to drive the display to display an endoluminal image of the lumen corresponding to the location.

There is further provided, in accordance with some applications of the present invention, a method for use with an input angiogram image of a device inserted inside a portion of a body of a subject, the device including radiopaque portions thereof, the angiogram image being acquired in the presence of contrast agent within the body portion, the method including:

generating, with a processor, a background image in which a relative value is assigned to a first pixel with respect to a second pixel, at least partially based upon relative values of surroundings of the first pixel and surroundings of the second pixel in the input image;

generating, with the processor, a cleaned image in which visibility of the radiopaque portions of the device is increased relative to the input image, by dividing the input image by the background image; and

generating an output on a display, based upon the cleaned image.

For some applications, generating the background image includes:

in the background image, assigning pixel values to the first pixel and the second pixel that are more similar to one another, than the similarity of pixel values that are assigned to a third pixel and a fourth pixel,

based upon the first pixel and the second pixel having more similar surroundings to one another in the input image, than a similarity of surroundings of the third pixel and the fourth pixel to one another in the input image.

For some applications, generating the background image further includes assigning values to the first and second pixels based upon values of the first pixel and the second pixel in the input image.

For some applications, generating the output includes displaying the cleaned image.

For some applications,

the input angiogram image of the device includes a plurality of input angiogram images of the device,

generating the cleaned image includes generating a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,

the method further includes generating a stabilized image stream by image tracking the cleaned images with respect to each other, based upon locations of the radiopaque portions of the device in the cleaned images, and

generating the output includes displaying the stabilized image stream.

For some applications,

the input angiogram image of the device includes a plurality of input angiogram images of the device,

generating the cleaned image includes generating a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,

the method further includes generating a stabilized image stream by image tracking the input images with respect to each other, based upon locations of the radiopaque portions of the device in the corresponding cleaned images, and

generating the output includes displaying the stabilized image stream.

For some applications,

the input angiogram image of the device includes a plurality of input angiogram images of the device,

generating the cleaned image includes generating a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,

the method further includes generating an enhanced image frame by:

• • aligning the cleaned images with each other, based upon locations of the radiopaque portions of the device in the cleaned images, and
  • generating an averaged image frame based upon the aligned cleaned images, and

generating the output includes displaying the enhanced image frame.

For some applications,

the input angiogram image of the device includes a plurality of input angiogram images of the device,

generating the cleaned image includes generating a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,

the method further includes generating an enhanced image frame by:

• • aligning the input images with each other, based upon locations of the radiopaque portions of the device in the corresponding cleaned images, and
  • generating an averaged image frame based upon the aligned input images, and

generating the output includes displaying the enhanced image frame.

For some applications,

the portion of the subject's body includes a lumen of the subject's body,

the device includes an endoluminal data-acquisition device configured to acquire endoluminal data points while the device is at respective locations within the lumen,

the method further includes, based upon locations of the radiopaque portions of the device in the cleaned image, determining that a given one of the endoluminal data points corresponds to a given location within the lumen, and

generating the output includes generating the output in response to the determining.

For some applications,

the portion of the subject's body includes a lumen of the subject's body,

the device includes an endoluminal device configured to be moved through the lumen,

the method further includes, based upon locations of the radiopaque portions of the device in the cleaned image, determining that the endoluminal device is at a given location within the lumen, and

generating the output includes, in response to the determining that the endoluminal device is at the given location within the lumen, generating an endoluminal image of the lumen corresponding to the location.

The present invention will be more fully understood from the following detailed description of embodiments thereof, taken together with the drawings, in which:

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A is a flow chart, at least some of the steps of which are used to generate a cleaned luminal image from an input luminal image, in accordance with some applications of the present invention;

FIG. 1B is a block diagram showing components of an image-processing system, in accordance with some applications of the present invention;

FIG. 2 shows an input luminal image, to which background cleaning may be subsequently applied, in accordance with some applications of the present invention;

FIG. 3 shows a background luminal image, generated in accordance with some applications of the present invention; and

FIG. 4 shows a clean luminal image, generated in accordance with some applications of the present invention.

DETAILED DESCRIPTION OF EMBODIMENTS

Terminology

• • The terms “medical tool,” “tool”, “device,” and “probe” refer to any type of a diagnostic or therapeutic or other functional tool including, but not limited to, a cardiovascular catheter, a stent delivery and/or placement and/or retrieval tool, a balloon delivery and/or placement and/or retrieval tool, a valve delivery and/or repair and/or placement and/or retrieval tool, a graft delivery and/or placement and/or retrieval tool, a tool for the delivery and/or placement and/or retrieval of an implantable device or of parts of such a device, an implantable device or parts thereof, a tool for closing a gap, a tool for closing a septal defect, a guide wire, a marker wire, a suturing tool, a clipping tool (such as a valve-leaflet-clipping tool), a biopsy tool, an aspiration tool, a navigational tool, a localization tool, a probe comprising one or more location sensors, a tissue characterization probe, a probe for the analysis of fluid, a measurement probe, an electrophysiological probe, a stimulation probe, an ablation tool, a tool for penetrating or opening partial or total occlusions in blood vessels, a drug or substance delivery tool, a chemotherapy tool, a photodynamic therapy tool, a brachytherapy tool, a local irradiation tool, a laser device, a tool for delivering energy, a tool for delivering markers or biomarkers, a tool for delivering biological glue, an irrigation device, a suction device, a ventilation device, a device for delivering and/or placing and/or retrieving a lead of an electrophysiological device, a lead of an electrophysiological device, a pacing device, a coronary sinus device, an imaging device, a sensing probe, a probe comprising an optical fiber, a robotic tool, a tool that is controlled remotely, or any combination thereof.
  • The terms “image” and “imaging” refer to any type of medical imaging, typically presented as a sequence of images and including, but not limited to, imaging using ionizing radiation, imaging using non-ionizing radiation, video, fluoroscopy, angiography, ultrasound, CT, MR, PET, PET-CT, CT angiography, SPECT, Gamma camera imaging, Optical Coherence Tomography (OCT), Near-Infra-Red Spectroscopy (NIRS), Vibration Response Imaging (VRI), Optical Imaging, infrared imaging, electrical mapping imaging, other forms of Functional Imaging, or any combination or fusion thereof. Examples of ultrasound imaging include Endo-Bronchial Ultrasound (EBUS), Trans-Thoracic Echo (TTE), Trans-Esophageal Echo (TEE), Intra-Vascular Ultrasound (IVUS), Intra-Cardiac Ultrasound (ICE), or any combination thereof.
  • The term “contrast agent,” when used in reference to its application in conjunction with imaging, refers to any substance that is used to highlight, and/or enhance in another manner, the anatomical structure, functioning, and/or composition of a bodily organ while the organ is being imaged.
  • The terms “lumen” and “cavity”, refer to any lumen or cavity of the cardiovascular system, the gastro-intestinal tract, the respiratory tract, the urinary tract, the nasal cavities, and any other bodily lumen or cavity.
  • The term “stabilized,” or “stable” when used in the context of displayed images, means a display of a series of images in a manner such that periodic, cyclical, and/or other motion of the body organ(s) being imaged, and/or of a medical tool being observed, is partially or fully reduced, with respect to the entire image frame, or at least a portion thereof.
  • The term “automatic,” when used for describing background cleaning, means “without necessitating user intervention or interaction.” (Such interaction or intervention may still however be optional in some cases.)
  • The term “real time” means without a noticeable delay.
  • The term “near real time” means with a short noticeable delay (such as approximately one or two motion cycles of an applicable organ, and, in the case of procedures relating to organs or vessels the motion of which are primarily as a result of the cardiac cycle, less than two seconds).
  • The term “on-line,” when used in reference to image processing, or to measurements being made on images, means that the image processing is performed, and/or the measurements are made, intra-procedurally, in real time or near real time.

References is now made to FIGS. 1A-1B, which are respectively (A) a flow chart, at least some of the steps of which are used to generate a cleaned luminal image from an input luminal image, in accordance with some applications of the present invention, and (B) a block diagram showing components of a processor of a an image-processing system, in accordance with some applications of the present invention. Applications of the present invention are typically used during medical procedures that are performed, in whole or in part, within body lumens or cavities. For some applications, apparatus and methods are provided for facilitating the visualization of a device positioned or deployed within such a lumen or cavity, at a time when the lumen or cavity is injected with contrast agent.

In accordance with some applications of the present invention, an input image acquired by an imaging device 10 (FIG. 1B) is input into a processor 11 of an image-processing system 12. Typically, within the input image, radiopaque portions of a tool inside a body lumen or a body cavity (and/or other are artifacts inside the lumen or the cavity) are hidden, in full or in part (e.g., are obscured), by contrast agent. The processor generates a cleaned image, in which visibility of the radiopaque portions of the tool (and/or the other artifacts) is improved relative to the input image. Typically, background-image-generation functionality 13 of the processor computes a background image. Subsequently, cleaned-image-generation functionality 14 of the processor divides the input image by the background image, such as to generate the cleaned image, in which visibility of the radiopaque portions of the tool (and/or the other artifacts) is improved relative to the input image. Typically, the cleaned image is generated from the input image automatically by the processor. Further typically, the generation of the cleaned image is performed by the processor on line, with respect to the inputting of the input image into the processor, and/or with respect to acquisition of the input image by an imaging device. Output generation functionality 15 of the processor typically drives a display 16 to display an output based on the cleaned image. For example, the display may display the cleaned image, or the display may display a stabilized image stream, the stabilization of the image stream being performed by performing image tracking based on locations of the radiopaque portions of the tool in the cleaned image.

Typically, in a first step, an input image u₀ (FIG. 1A) is inputted to processor 11, e.g., by the image being selected by the processor, or by a user. For some applications, the input image is a coronary angiogram, or a frame from an angiogram, that is acquired while a device that includes radiopaque sections thereof is within the coronary artery. For example, the device may be a balloon, a stent, a guide wire, an imaging probe, or any combination thereof. Typically, during injection of the contrast agent, the radiopaque sections of the device being inserted are hidden, in full or in part (e.g., obscured), by contrast agent.

Subsequent to the input image being inputted, a background image is generated by background-image-generation functionality 13, in accordance with the techniques described hereinbelow.

In the input image, pixels that are near to each other and that lie on the same object, are expected to have approximately the same value. For example, pixels lying on the ribs are expected to have approximately the same value as one another, and pixels lying inside a blood vessel are expected to have approximately the same value as one another. Thus, portions (i.e., pieces) of the input image are expected to have generally homogenous pixel values. However, the input image is not expected to be totally homogenous, since not all pixels lie on the same object. For example, pixels that lie on a rib are expected to have different values from pixels lying on a blood vessel. Thus, the values of the pixels in the input image can be assumed to be generally piecewise homogenous.

The assumption of piecewise homogeneity generally holds for the majority of the image pixels. However, the assumption fails to hold with respect to a portion of the pixels. An example of such pixels is that of the pixels that correspond to the radiopaque portions of an inserted device (e.g., radiopaque markers of a catheter). The values of such pixels are typically different from the values of their surrounding pixels, the surrounding pixels corresponding to the surrounding anatomy (e.g., a blood vessel in which the catheter is placed). Thus, these pixels are non-homogenous with respect to surrounding pixels.

For the purpose of the present invention, the non-homogeneous pixels are considered to be the foreground of the input image. An image that does not contain the non-homogeneous pixels (or in which the visibility of the non-homogeneous pixels is reduced), but which contains the piecewise homogenous pixels is considered to the background image. Thus, in accordance with the present invention, a background image is computed in which large contiguous piecewise homogenous image parts are enhanced relative to the input image, while the non-homogeneous pixels are made less visible relative to the input image. The background image is typically more homogeneous than the input image, and, in the background image, features of the subject's anatomy typically are enhanced relative to the input image, while the visibility of features of the tool is typically reduced relative to the input image.

The background image is typically computed by assigning a relative value to a first pixel in the background image with respect to a second pixel in the background image, based upon the relative values of the surroundings of the first pixel (e.g., a patch of pixels surrounding the first pixel) and the surroundings of the second pixel (e.g., a patch of pixels surrounding the first pixel) in the input image. Thus, pixels that have more similarly appearing surroundings in the input image are assigned more similar values to one another in the background image than pixels that have less similarly appearing surroundings in the input image. Typically, in computing the background image, the aforementioned method for assigning pixel values is traded-off against keeping the value of any given pixel in the background image similar to the value of the pixel in the input image.

For some applications, the background image is generated by computing a background image that is such as to reduce (e.g., that is such as to minimize) the cost of the following function:

$\begin{array}{cc}\sum _x\left[\frac{\lambda }{2}{\left(b\left(x\right)-u_0\left(x\right)\right)}^2+\sqrt{\sum _y{\left(b\left(y\right)-b\left(x\right)\right)}^{2}w\left(x,y\right)\mathrm{dy}}\right]& \left[\mathrm{Function}1\right]\end{array}$

in which:

• • b(x) is the value of pixel x is the background image;
  • u₀(x) is the value of pixel x in the input image;
  • b(y) is the value of pixel y is the background image;
  • w(x,y) is a weight measure that is the inverse of the weighted sum of the squared difference between p_x (i.e., a patch taken around pixel x) and p_y (i.e., a patch taken around pixel y).

The first term of Function 1 (i.e., the term in the brackets that appears before the “plus” sign) favors a background image whose pixel values are close to the pixel values of the input image. The second term of Function 1 (i.e., the term in the brackets that appears after the “plus” sign) favors a background image in which pixels the surroundings of which have similar values in the input image, have similar values in the background image. Thus, the resulting background image is typically similar to the input image, and at the same time is typically more homogenous than the input image, giving similar values to similar pixels. λ is a constant that represents the value of the trade-off between the first term and the second term of Function 1, i.e., the trade-off between (a) generating a background image in which the pixels have similar values to the values of the pixels in the input image and (b) generating an image in which the relative values of the pixels in the background image is based upon similarities between patches surrounding respective pixels in the input image. For some applications, the value of λ is set empirically, by testing different values on a wide range of benchmark input images.

For some applications, the background image is generated in accordance with minimization techniques described in “Nonlocal linear image regularization and supervised segmentation,” by Guy Gilboa and Stanley Osher (SIAM Multiscale Modeling & Simulation, volume 6, issue 2, pp. 595630, 2007), and/or in “Nonlocal Operators with Applications to Image Processing,” by Guy Gilboa and Stanley Osher (SIAM Multiscale Modeling & Simulation, volume 7, issue 3, pp. 1005-1028, 2008), both of which articles are incorporated herein by reference. Alternatively or additionally, other minimization techniques are used, such those described in “Non-local regularization of inverse problem,” by Gabriel Peyre, Sebastien Bougleux, and Laurent Cohenin (Lecture Notes in Computer Science, 2008, Volume 5304/2008, pp. 57-68), “The split Bregman method for L1 regularized problems,” by Tom Goldstein and Stanley Osher (SIAM Journal on Imaging Sciences, Volume 2, Issue 2, pp. 323-343), and/or in “Bregmanized nonlocal regularization for reconvolution and sparse reconstruction,” by Xiaoqun Zhang, Martin Burgery, Xavier Bresson, and Stanley Osher (SIAM Journal on Imaging Sciences, Volume 3, Issue 3, July 2010), all of which articles are incorporated herein by reference.

For some applications, an algorithm as described in “Bregmanized nonlocal regularization for reconvolution and sparse reconstruction,” by Zhang et al., which is incorporated herein by reference, is used to generate the background image. For example, algorithm 1 described on Page 17 of the aforementioned article may be used to generate the background image.

For some applications, the weight measure that is used to compute the background image (e.g., in accordance with Function 1, described hereinabove) is computed using the following technique. Given an input image, the processor of the system computes a weight measure between each pixel x in the image, and pixels in the vicinity of pixel x. The weight measure measures the similarity between small image patches centered on respective pixels x. For example, the similarity may be measured by an inverse of the weighted sum of squared differences between these patches. For some applications, the inverse of the weighted sum of squared differences is weighted by a Guassian, e.g., in accordance with techniques described in “A non-local algorithm for image denoising,” by Buades, Coll and Morell (IEEE CVPR 2005, volume 2, pages 60-65), which is incorporated herein by reference.

Typically, subsequent to the generation of the background image of the lumen, a cleaned image is generated by cleaned-image-generation functionality 14, by dividing the input image by the background image. (It should be noted that, throughout the description of the present invention, “dividing the input image by the background image” should be interpreted as being synonymous with “subtracting the background image from the input image” with respect to the mathematical operations that are performed by processor 11.) Typically, in the resulting cleaned image, image elements which are not homogeneous inside the lumen (such as the radiopaque markers of the device inserted in the vessel) remain visible while the vessel itself appears, in whole or in part, “clean” of contrast agent, at least relative to the input image.

Although some of the techniques described herein have been described with reference to an angiogram of a device that in inserted the into the coronary artery, the scope of the present invention includes applying the techniques to images of other body lumen and/or body cavities, mutatis mutandis. For example, the techniques described herein may be applied to an angiogram of the aorta that has an aortic replacement valve placed therein. Alternatively or additionally, the techniques described herein may be applied to an angiogram of a heart chamber that has a replacement valve placed therein. Further alternatively or additionally, the techniques described herein may be applied to an angiogram of a heart ventricle that has a ventricular assist device placed therein.

Reference is now made to FIG. 2, which shows an example of input luminal image. Specifically, the image shown in FIG. 2 is a frame of an angiogram of a coronary artery in which a catheter carrying a stent is inserted. Two radiopaque markers 20 that are disposed on the catheter may be observed at the proximal and distal edges of a balloon on which the pre-deployed stent is mounted.

Reference is now made to FIG. 3, which is a background luminal image generated from the input image, typically automatically and typically on line, in accordance with techniques described hereinabove. In FIG. 3 the contrast-filled coronary artery is clearly visible, while the radiopaque elements of the catheter carrying the stent are not visible.

Reference is now made to FIG. 4, which is a cleaned image that was produced by dividing the input image by the background image, in accordance with the techniques described hereinabove. The cleaned image was generated automatically and on line with respect to the inputting of the input image to the system. It may be observed that in the cleaned image, the contrast-filled vessel is largely invisible, while radiopaque markers 20 of the catheter carrying the stent are seen more clearly than in the input image.

For some applications, a series of cleaned image frames are used to create a stabilized image stream of an angiographic sequence, in which, typically, radiopaque elements of a device (e.g., markers of a catheter carrying a stent) appear relatively stable. Typically, the stabilized image stream of the angiographic sequence is generated in accordance with techniques described in U.S. patent application Ser. No. 12/650,605 to Cohen (published as US 2010/0172556) and/or in U.S. patent application Ser. No. 12/075,244 to Tolkowsky (published as US 2008/0221442), both of which applications are incorporated herein by reference. The stabilized image stream is typically displayed on display 16. For some applications, a plurality of cleaned images are generated, the cleaned images corresponding to respective input images. The input images are stabilized such as to generate a stabilized image stream, based upon the locations of the radiopaque elements of the device in the corresponding cleaned images. Alternatively or additionally, a cleaned, stabilized image stream is generated by stabilizing the cleaned images with respect to each other, based upon the locations of the radiopaque elements of the device in the cleaned images.

For some applications, a stent is deployed within a lumen, and a catheter and/or a balloon carrying radiopaque markers remains within the luminal section in which the stent is deployed. A series of image frames are cleaned in accordance with the techniques described herein, and the cleaned image frames are used to create an enhanced image stream of an angiographic sequence, in which the stent appears more visible than in a native angiogram of the stent. Typically, the enhanced image stream is generated in accordance with techniques described herein, in combination with techniques described in U.S. patent application Ser. No. 12/650,605 to Cohen (published as US 2010/0172556), which is incorporated herein by reference. The enhanced image stream is typically displayed on display 16. For some applications, a plurality of cleaned images are generated, the cleaned images corresponding to respective input images. An enhanced image frame is generated by aligning the input images with each other based upon locations of the radiopaque portions of the device in the corresponding cleaned images, and generating an averaged image frame based upon the aligned input images. Alternatively or additionally, a cleaned, enhanced image frame is generated, by aligning the cleaned images with each other, based upon locations of the radiopaque portions of the device in the cleaned images, and generating an averaged image frame based upon the aligned cleaned images.

For some applications, a series of cleaned image frames are used to create an image stream that is both stabilized and enhanced. Typically, the stabilized, enhanced image stream of the angiographic sequence is generated in accordance with techniques described herein, in combination with techniques described in U.S. patent application Ser. No. 12/650,605 to Cohen (published as US 2010/0172556) and/or in U.S. patent application Ser. No. 12/075,244 to Tolkowsky (published as US 2008/0221442), both of which applications are incorporated herein by reference. The stabilized, enhanced image stream is typically displayed on display 16.

For some applications, a series of input image frames are divided by the respective corresponding background images such as to produce cleaned image frames. The cleaned image frames are used to create an image stream that is stabilized, enhanced and cleaned, by stabilizing and enhancing the cleaned image frames. In an embodiment, such an image stream is produced in accordance with techniques described herein, in combination with techniques described in U.S. patent application Ser. No. 12/650,605 to Cohen (published as US 2010/0172556) and/or in U.S. patent application Ser. No. 12/075,244 to Tolkowsky (published as US 2008/0221442), both of which applications are incorporated herein by reference. The stabilized, enhanced, cleaned image stream is typically displayed on display 16.

For some applications, the visibility of the radiopaque markers on an endoluminal device is increased in an image stream, by cleaning image frames belonging to the image stream, in accordance with the techniques described herein. The increased visibility of the markers is used to facilitate tracking the device comprising those markers, typically automatically and typically on line. Typically, the tracking is performed in accordance with techniques described in U.S. patent application Ser. No. 12/650,605 to Cohen (published as US 2010/0172556), which is incorporated herein by reference. For some applications, the endoluminal device comprising the radiopaque markers is an endoluminal data-acquisition device (e.g., an endoluminal imaging probe), and the increased visibility of the radiopaque markers in the resulting image stream is utilized for co-registering, typically automatically and typically on line, endoluminal data points (e.g., endoluminal images) with the extraluminal images (e.g., extraluminal x-ray images). The endoluminal imaging probe may be ultrasound, optical coherence, infrared, MRI, or any combination thereof. Typically, the co-registration is performed in accordance with techniques described in International Patent Application PCT/IL2011/000612 (published as WO 12/014,212), which is incorporated herein by reference.

For example, based upon locations of radiopaque portions of an endoluminal data-acquisition device in the cleaned image the output-generation functionality of the processor may determine that a given one of the endoluminal data points corresponds to a given location within the lumen, and an output may be generated in response thereto. Alternatively or additionally, based upon locations of the radiopaque portions of an endoluminal device in the cleaned image, the output-generation functionality may determine that the endoluminal device is at a given location within the lumen. In response to determining that the endoluminal device is at the given location within the lumen, the output-generation functionality may drive the display to display an endoluminal image of the lumen corresponding to the location.

It will be appreciated by persons skilled in the art that the present invention is not limited to what has been particularly shown and described hereinabove. Rather, the scope of the present invention includes both combinations and subcombinations of the various features described hereinabove, as well as variations and modifications thereof that are not in the prior art, which would occur to persons skilled in the art upon reading the foregoing description.

Claims

1. Apparatus for use with at least one input angiogram image of a device inserted inside a portion of a body of a subject, the device including radiopaque portions thereof, the angiogram image being acquired in the presence of contrast agent within the body portion, the apparatus comprising: a display; andat least one processor configured: to generate a background image in which a relative pixel value is assigned to a first pixel with respect to a second pixel, at least partially based upon relative pixel values of a patch of pixels that surround a first pixel of the input angiogram image that is disposed at the same location as the first pixel of the background image and a patch of pixels that surround a second pixel of the input angiogram image that is disposed at the same location as the second pixel of the background image;to generate a cleaned image in which visibility of the radiopaque portions of the device is increased relative to the input angiogram image, by dividing the input angiogram image by the background image; andto drive the display to display an output based upon the cleaned image.

2. The apparatus according to claim 1, wherein the processor is configured to generate the background image by: in the background image, assigning pixel values to the first pixel and the second pixel that are more similar to one another, than the similarity of pixel values that are assigned to a third pixel and a fourth pixel,based upon the patch of pixels that surround the first pixel of the input angiogram image and the patch of pixels that surround the second pixel of the input angiogram image having more similar pixel values to one another, than a similarity, to one another, of pixel values of a patch of pixels that surround a third pixel of the input angiogram image that is disposed at the same location as the third pixel of the background image and a patch of pixels that surround a fourth pixel of the input angiogram image that is disposed at the same location as the fourth pixel of the background image.

3. The apparatus according to claim 1, wherein the processor is further configured to generate the background image by assigning pixel values to the first and second pixels of the background image based upon pixel values of the first pixel and the second pixel of the input angiogram image.

4. The apparatus according to claim 1, wherein the processor is configured to drive the display to display the cleaned image.

5. The apparatus according to claim 1, wherein: the input angiogram image of the device includes a plurality of input angiogram images of the device, andthe processor is configured to: generate a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,generate a stabilized image stream by image tracking the cleaned images with respect to each other, based upon locations of the radiopaque portions of the device in the cleaned images, anddrive the display to display the stabilized image stream.

6. The apparatus according to claim 1, wherein: the input angiogram image of the device includes a plurality of input angiogram images of the device, andthe processor is configured to: generate a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,generate a stabilized image stream by image tracking the input images with respect to each other, based upon locations of the radiopaque portions of the device in the corresponding cleaned images, anddrive the display to display the stabilized image stream.

7. The apparatus according to claim 1, wherein: the input angiogram image of the device includes a plurality of input angiogram images of the device, andthe processor is configured to: generate a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,generate an enhanced image frame by: aligning the cleaned images with each other, based upon locations of the radiopaque portions of the device in the cleaned images, andgenerating an averaged image frame based upon the aligned images, anddrive the display to display the enhanced image frame.

8. The apparatus according to claim 1, wherein: the input angiogram image of the device includes a plurality of input angiogram images of the device, andthe processor is configured to: generate a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,generate an enhanced image frame by: aligning the input images with each other, based upon locations of the radiopaque portions of the device in the corresponding cleaned images, andgenerating an averaged image frame based upon the aligned images, anddrive the display to display the enhanced image frame.

9. The apparatus according to claim 1, wherein: the portion of the subject's body includes a lumen of the subject's body,the device includes an endoluminal data-acquisition device configured to acquire endoluminal data points while the device is at respective locations within the lumen, andthe processor is configured: based upon locations of the radiopaque portions of the device in the cleaned image, to determine that a given one of the endoluminal data points corresponds to a given location within the lumen, andto drive the display to display an output, in response to the determining.

10. The apparatus according to claim 1, wherein: the portion of the subject's body includes a lumen of the subject's body,the device includes an endoluminal device configured to be moved through the lumen, andthe processor is configured: based upon locations of the radiopaque portions of the device in the cleaned image, to determine that the endoluminal device is at a given location within the lumen, andin response to the determining that the endoluminal device is at the given location within the lumen, to drive the display to display an endoluminal image of the lumen corresponding to the location.

11. A method for use with at least one input angiogram image of a device inserted inside a portion of a body of a subject, the device including radiopaque portions thereof, the angiogram image being acquired in the presence of contrast agent within the body portion, the method comprising: generating, with a processor, a background image in which a relative pixel value is assigned to a first pixel with respect to a second pixel, at least partially based upon relative pixel values of a patch of pixels that surround a first pixel of the input angiogram image that is disposed at the same location as the first pixel of the background image and a patch of pixels that surround a second pixel of the input angiogram image that is disposed at the same location as the second pixel of the background image;generating, with the processor, a cleaned image in which visibility of the radiopaque portions of the device is increased relative to the input angiogram image, by dividing the input angiogram image by the background image; andgenerating an output on a display, based upon the cleaned image.

12. The method according to claim 11, wherein generating the background image comprises: in the background image, assigning pixel values to the first pixel and the second pixel that are more similar to one another, than the similarity of pixel values that are assigned to a third pixel and a fourth pixel,based upon the patch of pixels that surround the first pixel of the input angiogram image and the patch of pixels that surround the second pixel of the input angiogram image having more similar pixel values to one another, than a similarity, to one another, of pixel values of a patch of pixels that surround a third pixel of the input angiogram image that is disposed at the same location as the third pixel of the background image and a patch of pixels that surround a fourth pixel of the input angiogram image that is disposed at the same location as the fourth pixel of the background image.

13. The method according to claim 11, wherein generating the background image further comprises assigning pixel values to the first and second pixels of the background image based upon pixel values of the first pixel and the second pixel of the input angiogram image.

14. The method according to claim 11, wherein generating the output comprises displaying the cleaned image.

15. The method according to claim 11, wherein the input angiogram image of the device includes a plurality of input angiogram images of the device,wherein generating the cleaned image comprises generating a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,the method further comprising generating a stabilized image stream by image tracking the cleaned images with respect to each other, based upon locations of the radiopaque portions of the device in the cleaned images, andwherein generating the output comprises displaying the stabilized image stream.

16. The method according to claim 11, wherein the input angiogram image of the device includes a plurality of input angiogram images of the device,wherein generating the cleaned image comprises generating a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,the method further comprising generating a stabilized image stream by image tracking the input images with respect to each other, based upon locations of the radiopaque portions of the device in the corresponding cleaned images, andwherein generating the output comprises displaying the stabilized image stream.

17. The method according to claim 11, wherein the input angiogram image of the device includes a plurality of input angiogram images of the device,wherein generating the cleaned image comprises generating a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,the method further comprising generating an enhanced image frame by: aligning the cleaned images with each other, based upon locations of the radiopaque portions of the device in the cleaned images, andgenerating an averaged image frame based upon the aligned cleaned images, andwherein generating the output comprises displaying the enhanced image frame.

18. The method according to claim 11, wherein the input angiogram image of the device includes a plurality of input angiogram images of the device,wherein generating the cleaned image comprises generating a plurality of cleaned images, the cleaned images corresponding to respective input angiogram images,the method further comprising generating an enhanced image frame by: aligning the input images with each other, based upon locations of the radiopaque portions of the device in the corresponding cleaned images, andgenerating an averaged image frame based upon the aligned input images, andwherein generating the output comprises displaying the enhanced image frame.

19. The method according to claim 11, wherein the portion of the subject's body includes a lumen of the subject's body,wherein the device includes an endoluminal data-acquisition device configured to acquire endoluminal data points while the device is at respective locations within the lumen,the method further comprising, based upon locations of the radiopaque portions of the device in the cleaned image, determining that a given one of the endoluminal data points corresponds to a given location within the lumen, andwherein generating the output comprises generating the output in response to the determining.

20. The method according to claim 11, wherein the portion of the subject's body includes a lumen of the subject's body,wherein the device includes an endoluminal device configured to be moved through the lumen,the method further comprising, based upon locations of the radiopaque portions of the device in the cleaned image, determining that the endoluminal device is at a given location within the lumen, andwherein generating the output comprises, in response to the determining that the endoluminal device is at the given location within the lumen, generating an endoluminal image of the lumen corresponding to the location.