Patent Application
20090046192
The present invention relates to an automated digital image recording system and a method accomplished by means of such a system for digitizing slides, wherein said image recording system comprises a slide holding means, a mechanism for actuating the slide holding means, a first digital image recording means, a high magnification second digital image recording means and a control unit for software-based control of the automated image capturing.
Evaluating tissue biopsies is of high importance in the field of medical diagnostics. For performing histological evaluation, a piece of tissue is taken from a patient and sliced up into very thin sections. Then the sections are placed on glass plates and stained, and the slide with the thus obtained sample is sub-jected to a microscopic examination.
The efficiency of the evaluation can be significantly increased by first digitizing the samples with an automated imaging microscope and afterwards evaluating the digital images obtained in this way by an image managing routine. By most digital microscopes used nowadays, the slides are digitized in a very large number of steps, that is frame by frame, wherein each frame corresponds to a single microscopic field of view. The slide is arranged on a stage governed by a computer that controls the microscope, as well as the focusing on the sample.
The greater number of slides should be digitized, the more important it becomes to decrease the total operation time per slide. The total operation time is affected by the following major factors: the number of frames (or fields of view) to be imaged, the time taken by the starting, moving and stopping of the stage, the focusing time spent on the individual fields of view, and the time for completing the operations associated with the digital image capturing itself. Apparently, a decrease in the number of frames to be read leads to an automatic decrease in all further factors. Therefore, an aim of the present invention is to reduce the number of frames to be read at high magnification.
The microscope according to U.S. Pat. No. 6,101,265 is equipped with two objectives, one of a smaller and one of a higher magnification. First, the slide is imaged as field of view by making use of the smaller magnification objective and the digital frames obtained in this way are then matched together and displayed on a screen. At this point, the pathologist carrying out the examination can choose the region of interest for him/her, which is then scanned again, at this time by making use of the higher magnification objective, and after having matched the frames together, the chosen digitized slide region is shown again to the user. Hence, here the number of frames to be imaged at high magnification is reduced only in case of a real-time evaluation, and thus the problem of fast and automated imaging of a great number of slides at high magnification remains to be unsolved. Moreover, the two microscopic objectives considerably raise the cost of such a device.
U.S. Patent Appl. No. 2002/0090127, being the closest prior art according to our knowledge, discloses such a solution wherein a digital preview image of the slide as a whole is captured by a preview camera. The preview image is processed by a computer and based on a characteristic value (e.g. the light intensity) of each individual pixel the sample's position on the slide is determined. Then, only the fields of view corresponding to these pixels of the sample are imaged with the high magnification microscope. A drawback of the solution is that it is inappropriate for imaging and processing of other data/pieces of information being also on a slide, or rather if there are such further pieces of information present, they are either also scanned—superfluously—at high magnification since based on a simple study of the pixels they seem to belong to the sample, or a complicated additional routine is required that is suitable for discriminating the sample and these further pieces of information from one another. For instance, by means of a simple study of the light intensity, the legends or bar codes would be classified as areas to be also scanned, and therefore to avoid this, sophisticated expert systems are applied which take the sample's typical arrangement and morphology also into account when choosing the fields of view to be scanned by the high magnification microscope. Due to the interim analysis being completed, this, however, increases the duration of image capturing and/or the computing capacity required.
Such further pieces of information can be, for instance, the supplementary data indicated on the slide with a tissue sample used customarily when evaluating tissue samples, which e.g. can serve for identifying the sample. Typically, the pieces of information are provided in a region located in the vicinity of the end of the slide where no sample is present, e.g. in a free region situated at one end of the slide. The pieces of information can be provided in the form a handwriting, however, by an increase in the number of samples to be studied, the application of a bar code would be more and more preferred for the identification of the samples.
In order to reduce the number of frames to be read by the high magnification objective and simultaneously to decrease the total operation time of digital recording per slide, as well as to eliminate the above-identified problems, an automated digital image recording system is provided that serves for digitizing a slide having one or more regions with a sample and at least one separate region with a field of information. The first digital image recording means is adapted for reading out the at least one region containing the field of information. The image recording means comprises a preview camera for capturing a separate digital preview image of each sample containing region. The control unit is provided with a routine for searching pixels belonging to the samples in the preview images.
Preferred embodiments of the digital image recording system according to the invention are defined by claims 2 to 9.
Within the framework of the method according to the invention,
(a) a slide having one or more regions with a sample and at least one separate region with a field of information is arranged within the slide holding means;
(b) the slide is moved to the first digital image recording means equipped with a preview camera by the actuating mechanism;
(c) the at least one region with the field of information is read out by the first digital image recording means and a separate digital preview image of each of the one or more regions with a sample is captured by the preview camera;
(d) pixels belonging to the sample are detected in the preview image(s);
(e) the slide is moved to the second digital image recording means by the actuating mechanism and a high magnification digital image of all the fields of view (M) that correspond to the pixels detected is captured.
Possible further preferred variants of the method according to the invention are defined by claims 11 to 22.
Further details of the invention are discussed in relation to various embodiments and referring to drawings, wherein
Between the preview camera 2 and the objective 3 a slide 6 (preferentially a microscope slide with a biological/tissue sample arranged thereon) fixed in a slide holding frame 7 is moved by means of a slide displacing mechanism. Such a slide displacing mechanism is disclosed e.g. by International Patent Appl. No. PCT/IB2005/050351, and hence it is not discussed here in more detail. The slide holding frame 7 is attached to a stage 17 through a known type of suspension, wherein the stage 17 is affixed to a threaded bar 9 driven by a stepping motor 8 capable of effecting displacement along the axis X. Here, the threaded bar 9 brings about displacements of the stage 17 along a rail 10 lying in the axis X and, in turn, through the suspension, that of the slide holding frame 7 and the slide 6. Similarly, the rail 10 is affixed to a rail 12 lying in the axis Z of a stepping motor 11 that is capable of bringing about displacements along the axis Z. The rail 10 can be moved by means of a threaded bar 13 driven by the stepping motor 11. Thus, both the slide holding frame 7 and the slide 6 can be moved in both the X and Z directions. Here, the control of the stepping motors 8, 11 is equally performed by the control unit provided by the computer 4 in the present case.
The plane spanned by the axes X, Z can also be a horizontal one; in this case it is adequate if the slide 6 is supported by the slide holding frame 7 from below. In what follows, however, such an embodiment is shown, wherein the axis Z is vertical and hence the slide 6 is arranged in an edgewise position (that is, when slides of a rectangular shape are used, as is the case in general in the field of medical diagnostics, preferably laid flatwise on its longer edge) within the slide holding frame 7. It is noted, that the plane spanned by the axes X, Z can be arranged actually at any angle relative to the horizontal.
Preferably, a slide dispensing means 14 is connected to an end of the rail 10 parallel to the X axis that loads (unloads) the slides 6 into (from) the slide holding frame 7 automatically. Such a slide dispensing means 14 is disclosed e.g. by International Publication Pamphlet No. WO2004/113989.
The illumination required for the image capturing is preferably provided on the one hand by a light source 15 arranged in the vicinity of the preview camera 2 and directed onto the slide 6, and on the other hand by a light source 16 arranged on a side of the slide 6 opposite the objective 3. Preferably, both light sources 15, 16 are connected to the computer 4. Here, transmitted light microscopy is discussed, however, in case of other kinds of microscopy the illumination of the slides 6 can be realized in different manners; optionally, the light source(s) 15 and/or 16 might be dispensable and, hence, can even be omitted.
Since four times the area of the field of view E of the preview camera 2 is larger than the total area of the actual regions 18, 19, 20, 21, it is preferred to crop the image of the portions located outside the slide 6 from the image of the field of view E of the preview camera 2. As it is obvious for a person skilled in the relevant art, this can be easily achieved by the application of a suitable software means.
The first region 18 is the field of information of the slide 6; the data/pieces of information for identifying a sample 22 present on the slide 6, optionally any further notations carrying information, can be found here. The field of information may contain, for instance, a legend 23 comprising printed or handwritten letters and/or numbers, or a bar code 24, or a further piece of visual information applied onto the slide, or any combination thereof.
As the field of information requires a different way of processing, it is preferred to deal with it differently as compared to the other regions 19, 20, 21 with a sample 22. In case of a legend 23, the computer 4 can be equipped with an optical character recognition, in particular with a writing recognition routine. The text recognized can be stored as a text file together with the high magnification image files of the slide 6, or can be optionally used to define the filenames of the image files. If bar codes 24 are used, the computer 4 can be equipped with a bar code recognition/reading routine and the (textual) information deciphered can be saved together with the high magnification image files of the slide 6. The advantage of this latter embodiment is that, on the one hand, by using the bar code 24, compared to handwriting, much more information can be included in the field of information, and on the other hand it is significantly cheaper to read the bar code 24 by means of a web camera and a bar code recognition/reading routine than to incorporate a complete bar code recognition/reading equipment into the system 1. Furthermore, an embodiment is also possible wherein the preview camera 2 is used merely to scan the regions 19, 20, 21 with a sample 22, while the region 18 with the field of information is read and processed by a bar code reader. In a further possible embodiment (if handwriting is contained in the field of information), the field of information is simply digitized by the preview camera 2, the image thus scanned is stored as an image file and shown to a user on demand who interprets the pieces of information that can be seen in the image himself/herself.
The embodiment of the image recording system 1 according to the present invention shown in
The slides 6 to be digitized are arranged in the slide dispensing means 14 that loads the first slide 6 into the slide holding frame 7 by means of e.g. a robotic arm. If the image recording system 1 is not equipped with a slide dispensing means 14, the slide 6 is loaded into the image recording system 1 by the operator through an opening formed for this purpose, where the slide 6 goes into the slide holding frame 7. Upon control of the computer 4 the following operations are carried out. The stepping motor 8 moves the slide holding frame 7 along the rail 10 parallel to the axis X in front of the preview camera 2 so as to locate the first region 18 within the field of view E of the preview camera 2. A digital image of region 18 is captured, then the slide holding frame 7 is slided along the rail 10 by the stepping motor 8 so as to locate the second region 19 within the field of view E of the preview camera 2 and then a preview image of this region 19 is also captured. This procedure is then continued in a similar manner until a preview of the last region 21 is captured.
The image of the first region 18 is processed according to the nature of the information found in the field of information by one of the procedures already mentioned, and is stored in the computer 4 or in a suitable data storage means connected to the computer 4 through a network or other ways. Processing of the region 18 can take place simultaneously with capturing the preview images of the other regions 19, 20, 21. Indeed, the image recording system 1, as well as the slide dispensing means 14 can also be formed in such a manner that the region 18 with the field of information behind, that is, when the slide 6 is moved towards the preview camera 2 along the rail 10, at first region 21, then region 20, then region 19 and finally region 18 reaches the preview camera 2. In this case the recording can be commenced by region 21 and the preview images of regions 20, 19, 18 can be captured by displacing further the slide 6 along the axis X. A further possibility is to bring directly the region 18 in front of the preview camera 2 and then to capture regions 19, 20, 21 progressing backwards from it. Optionally, the computer 4 might be equipped with an image recognition routine being capable of making a decision whether the region 18, 19, 20 or 21 at issue contains a field of information or a portion of the sample 22, and the digital image is processed accordingly.
The digital preview images of the regions 19, 20, 21 are used for a software-based search of the location of the sample 22. This means that pixels belonging to the sample 22 are being searched within the regions 19, 20, 21 with a sample 22 of the digitized slide 6, and their spatial coordinates relative to the stage 17 within a reference frame defined by the stepping motors 8, 11 are determined. To the pixels classified as belonging to the sample 22, respective portions of the field of view M of the second image recording means (i.e. of the digital microscope) are assigned (preferably in an overlapping manner) and as a next step the thus defined fields of view M containing each a portion of the sample 22 are digitized at a high resolution.
To decide which pixel of the preview image belongs to the sample 22 (that is, whether the field of view M assigned to the pixel at issue during the high resolution and high magnification digitizing contains a portion of the sample 22), there are numerous procedures known from the fields of image recognition and image processing. A simplest solutions is to consider the light intensity of pixels, and if a certain threshold value thereof is reached, taking the pixel as belonging to the sample 22 and then defining to which field(s) of view M the pixel considered belongs to.
Furthermore, to the fields of view M containing the sample 22, X, Y coordinates are assigned and it is calculated into which positions the stage 17 should be moved in order that the field of view M concerned could be captured by the high magnification digital microscope.
For example, to perform a high resolution and high magnification scanning a 20× microscope objective 3 and a CCD camera 5 with the resolution of 1024 pixels by 768 pixels are used. Before the 20× magnification, an area with the size of about 0.23 μm by 0.23 μm of the slide 6 corresponds to a single pixel of the digital image captured by the second image recording means comprising the objective 3 and the CCD camera 5. This means that an area with the size of about 235 μm by 176 μm of the slide 6 could be digitized at each shot, i.e. this is the size of the field of view M of the second image recording means. It is noted here that at least an 5× microscope objective should be preferably used to end up with digital images that are of suitable quality for practical usage.
Before the next field of view M containing a portion of the sample 22 could be digitized, the captured image must be downloaded to the computer 4. In one of the embodiments this takes generally about 65 ms, depending on the size of the image and the quality of the connection. This is longer than the time period required for the slide 6 to be moved into its next position. After exposition, the displacement of the stage 17 immediately commences, thus by the time the image has been downloaded, the slide 6 already occupies a new position to be scanned. For the sake of simplicity, the digital images and frames created via scanning are referred to by the adjective “digital”, that is, from now on the concepts of digital field of view M, digital sample 22 and digital slide 6 are used.
After all the fields of view M chosen on basis of the preview image have been scanned at high resolution and high magnification, or simultaneously with the scanning, the digital fields of view M are matched along their adjacent edges (that can be optionally overlapping). A first possibility is to store the digital slide 6 containing the digital sample 22 composed of the digital fields of view M that are matched together. A further possibility is to store merely the data related to the matchings themselves (contiguity relations, overlappings of individual frames, measures of translation, etc.) along with the files of the individual digital fields of view M, and when the digital slide 6 is displayed the files of the digital fields of view M are displayed in harmony with the data related to the matchings. A half-way measure is also conceivable; a given number of neighbouring digital fields of view M is matched together in advance resulting in tiny digital fields, and the tiny fields as well as the data related to the matchings thereof are stored in files. In other cases only the files of the digital fields of view M are stored, and later on put together and looked at by means of a suitable digital slide displaying and manipulating routine.
After high resolution and high magnification scanning of the slide 6, the slide holding frame 7 is returned to the slide dispensing means 14 by means of the stepping motors 8, 11, where the already scanned slide 6 is replaced via the robotic arm by a further slide 6 to be also scanned. In lack of a slide dispensing means 14, the exchange of slides 6 can be effected manually or via other means.
In certain embodiments the computer 4 is connected to a local network or to the Internet. In this case storage, display and further operations do not take place definitely in the computer 4. Indeed, there is no need for the computer 4 performing control of the image recording system 1 and the image recording system 1 itself to be physically at the same location; the control itself can also be realized through a network.
In this latter case the preview camera 2 can be arranged so as to capture the region 26 with the sample 22 of the slide 6 as a single frame. Indeed, the slide 6 can be divided into more than one sample containing regions 19, 20, 21 in this case, too.
After scanning the bar code 24 and capturing the preview image, the operation is identical to that of the first embodiment.
The embodiments shown above are exemplary only; it is clear that a person skilled in the relevant field can effect a number of changes without departing the scope of protection defined by the attached set of claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| P0600177 | Mar 2006 | HU | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/HU07/00018 | 3/5/2007 | WO | 00 | 10/6/2008 |
