JP6931579B2 - Live cell detection methods, programs and recording media - Google Patents

Description

The present invention relates to a technique for identifying a living cell, cell mass, or the like included in an image.

In fields such as drug discovery screening and biochemical experiments, images obtained by imaging cells cultured in a culture vessel or a cell mass consisting of a plurality of cells (hereinafter, these are collectively referred to as "cells, etc."). Is subjected to appropriate image processing to evaluate the state of cells and the like. In particular, since a visible light image obtained by imaging unstained cells or the like under visible light illumination can be imaged without damaging living cells or the like, various information should be obtained from such an image. Is desired. On the other hand, it is hard to say that the evaluation criteria of the image are sufficiently established as compared with the image taken by using a reagent (for example, staining or fluorescence labeling) that is expressed only in cells having a specific attribute. .. Therefore, various techniques for image evaluation have been proposed.

For example, in the technique described in Patent Document 1, in a visible light image obtained by imaging unstained cells, a region having a brightness higher than an appropriately set threshold value is formed based on the finding that the brightness of living cells is higher than that of dead cells. A region corresponding to a living cell and a region having a lower brightness are determined to be dead cells or a background region. Such a process of determining the life or death of a cell or the like is an effective means for evaluating, for example, the medicinal effect of a chemical substance on a cell.

Japanese Unexamined Patent Publication No. 2016-116461

As described in Patent Document 1, in general, living cells tend to show higher brightness than dead cells in images. However, such a difference in brightness is relative, and the brightness in the image varies greatly due to, for example, variations in the color of individual cell bodies and differences in lighting conditions. For this reason, it may not be possible to accurately detect living cells or the like by a determination based only on an absolute brightness value such as a comparison with a mere threshold value.

The present invention has been made in view of the above problems, and provides a technique capable of detecting a living object with higher accuracy than before from an image including a cell, a cell mass, etc. as a detection target. The purpose.

One aspect of the live cell detection method according to the present invention comprises a step of acquiring an original image including a visible light image of a cell or a cell mass as a detection target and the original image in order to achieve the above object. For each pixel, a step of executing a minimum value filter process based on the brightness value of the pixel and a nearby pixel within a predetermined neighborhood range centered on the pixel, and each pixel constituting the image after the minimum value filter process. On the other hand, the step of executing the maximum value filtering process based on the brightness value of the pixel and the neighboring pixel in the neighborhood range having the same size as the neighborhood range in the minimum value filtering process centered on the pixel, and the original image and the said It includes a step of obtaining a difference from the image after the maximum value filtering process, and a step of specifying a region occupied by pixels having a brightness value equal to or higher than a predetermined threshold in the difference as a region corresponding to a living cell.

Although details will be described later, the present invention is based on the following findings. That is, in the image, the peripheral portion of a cell or a cell mass composed of a plurality of cells (hereinafter, collectively referred to as "cells and the like") exhibits a relatively low brightness with respect to the background. On the other hand, in the central portion of cells and the like, living cells and the like exhibit relatively high brightness with respect to the peripheral portion, whereas dead cells and the like have low brightness as in the peripheral portion.

In the present invention, the minimum value filtering process is performed on the original image including cells and the like. In the minimum value filtering process, the brightness value of each pixel constituting the image is replaced by the minimum value of the brightness value of the pixel and its surrounding neighboring pixels. As a result, in the image of a living cell or the like, the brightness value of the pixel in the central portion having a relatively high brightness is replaced with the brightness value of the pixel in the peripheral portion having a lower brightness, and the entire inside of the contour is darkly filled. Become. On the other hand, in the image of dead cells and the like, since the brightness in the central part is originally low, the change due to the treatment is small. Further, even in the background portion where the change in brightness is relatively small, the change due to processing is small.

By executing the minimum value filtering process, a phenomenon that the contour expands in the peripheral portion of a cell or the like inevitably occurs. By executing the maximum value filtering in the neighborhood range of the same size as the neighborhood range applied to the minimum value filtering, the expanded contour can be brought closer to the original shape. The central part, which is already darkly filled in the image of living cells, does not return to the original high brightness value even by the maximum value filtering. Therefore, in the image of living cells and the like in the image at this time, the central portion is darkly filled while maintaining the original contour shape. On the other hand, in dead cells and the like, the change in brightness is originally small, so the change due to treatment is small.

When the difference obtained by subtracting the image obtained in this way from the original image is obtained, the region where the brightness changes significantly, that is, the region corresponding to the central part such as a living cell shows a high brightness value, while the dead cell with a small change in brightness shows a high brightness value. Areas such as and the background portion show sufficiently low brightness. In this way, in the difference image, living cells and the like and other parts appear as a large difference in brightness. Therefore, it can be said that the region having a certain brightness value or more in the difference image indicates the position where the living cells or the like existed in the original image. From this, the region occupied by the pixels having the brightness value equal to or higher than the predetermined threshold value in the difference image can be regarded as the region corresponding to the living cells and the like.

As described above, according to the present invention, such a region is specified as a region corresponding to the living cell by utilizing the characteristic of the living cell that the brightness of the central portion is higher than that of the peripheral portion. Therefore, it is possible to detect a living object from an image including a cell body such as a cell or a cell mass with higher accuracy than before.

It is a schematic diagram which shows the example of the original image. It is a flowchart which shows the processing content of one Embodiment of the living cell detection method which concerns on this invention. It is a figure which shows the example of the image represented by the image data generated in the process. It is a figure which illustrates the effect of the process of this embodiment. It is a figure which illustrates the window size in the filtering process. It is a figure which shows another example of image processing. It is a figure which shows another example of image processing.

Hereinafter, the principle and specific processing contents of the living cell detection method according to the present invention will be described. Hereinafter, the term "cell mass" is used as a general term for organoids (cell tissues) and spheroids that form a cohesive structure in which a plurality of cells are aggregated. In addition, the term "cells and the like" is used as a general term including a cell mass and a cell existing alone. In addition, "living" cell mass means that the main part of the cell mass is composed of active cells. On the other hand, a "dead" cell mass means a state in which all or most of the cells constituting the cell mass have died or lost their activity. In addition, "living cell" is used as a general term for a cell mass composed of a single living cell and a plurality of living cells.

The living cell detection method according to the present invention uses an image obtained by imaging a cell or the like as an original image, and detects a living cell or the like in the original image using the cell or the like contained in the original image as a detection target. It is a thing. As the detection target, for example, cells or cell clusters cultured in a medium injected into a culture vessel such as a well plate or a dish can be used. Further, as the original image, an image taken by a bright-field microscope under, for example, visible light illumination in the culture vessel containing such a detection object can be used.

FIG. 1 is a schematic diagram showing an example of an original image. The original image Is is an image obtained by bright-field imaging of the inside of a culture vessel containing, for example, unstained cells cultured in a medium. In the original image Is, an image of at least one cell or the like is included in the background BG having substantially uniform brightness corresponding to the medium. The original image Is shown in FIG. 1 includes a plurality of images of organoids (cell tissues) formed by aggregating a plurality of cells. As is traditionally known, images of living organoids are relatively bright, and dead organoids appear in the image as dark images. Here, a case where an organoid is targeted as an example of "cells and the like" will be described, but the application target of the treatment is not limited to the organoids, but the cells and the like in general.

By the way, when the organoids in the image are observed in more detail, as shown by the reference numerals C1 to C4 in FIG. 1, the living organoids have higher brightness in the central portion and the peripheral portions have higher brightness than the central portion. Is low. On the other hand, the dead organoids have a uniformly low brightness and a dark image in the entire image as shown by the reference numerals C5 to C10. In addition, the brightness of each organoid also differs from individual to individual between living organoids C1 to C4 and between dead organoids C5 to C10. Such variations in the brightness value also occur due to differences in the imaging environment such as lighting conditions and background brightness. For this reason, it is not always possible to make a highly accurate determination by a method of determining whether an organoid appearing in an image is alive or dead based solely on its overall brightness value.

On the other hand, as described above, the living organoid has a property that the brightness of the peripheral portion is low and the brightness of the central portion is higher than that of the peripheral portion. By utilizing this property, it is possible in principle to accurately detect living organoids without being affected by variations depending on the individual and the imaging environment. The living cell detection method according to the present invention is based on this finding.

FIG. 2 is a flowchart showing the processing content of one embodiment of the living cell detection method according to the present invention. Further, FIG. 3 is a diagram showing an example of an image represented by image data generated during processing. This processing can be realized, for example, by causing a computer device having a hardware configuration general as a personal computer to execute a program describing the processing procedure described below. For the sake of explanation, FIG. 3 also includes an image corresponding to a visualization of image data intermediately generated in the middle of processing, but in actual processing, an image corresponding to such intermediate data is included. It is not mandatory to be created.

First, image data (original image data) corresponding to the original image to be processed is acquired (step S101). Here, as shown in FIG. 3, the living organoids Ca and Cb whose central portion is brighter than the peripheral portion and the dead organoids Cc and Cd whose central portion and peripheral portion have approximately the same brightness are included. Image I0 is used as the original image. The original image can be obtained by operating the imaging device and imaging the cells and the like in the culture vessel. In this case, this process may be executed by the processor included in the image pickup apparatus. In addition, for example, it is possible to acquire image data by reading the image data created by imaging in advance from a recording medium on which the image data is recorded, or by receiving the image data from another computer device via a communication line. be.

The minimum value filtering process (step S103) and the maximum value filtering process (step S104) are sequentially executed for each pixel constituting the acquired original image I0, but prior to that, the window applied to the filtering process is displayed. The size is set (step S102). The window size in the minimum value filtering process and the maximum value filtering process determines the neighborhood range with respect to the central pixel to be processed, and the effect of the filtering process differs depending on the size setting. Therefore, it is necessary to set an appropriate window size according to the content of the image, specifically, according to the size of the cells and the like included in the original image I0. The preferred window size will be described later.

If the size of the cell or the like is known in advance, the window size can also be fixedly determined in advance. If the type of cells or the like (for example, organoids) to be observed is determined, the degree of variation in size can be predicted in advance. Further, when the original image I0 is imaged, if the magnification is set so that the cells and the like in the image have a size suitable for observation, the size can be substantially the same as the size in the image regardless of the type of the cells and the like. be. Therefore, such a mode in which the window size is predetermined is sufficiently practical.

On the other hand, it is also conceivable to make the window size variable in order to obtain better processing results. In this case, for example, the window size may be set based on the information given by the instruction input from the user, and the window size is determined from the size of the object corresponding to the cells or the like in the image by an appropriate image analysis process. May be good.

Based on the window size set in this way, the minimum value filtering process is executed for the original image I0 (step S103). The minimum value filter process is a process of replacing the value (luminance value) of the center pixel, that is, the pixel to be processed located at the center of the window, with the minimum value of the brightness value of each pixel in the window. By this process, the value of a bright pixel having a high luminance value is replaced with the luminance value of a darker pixel in the vicinity thereof.

When the processed image data is visualized as an image when the window size is set appropriately, as shown in image I1 shown in FIG. 3, the bright portion in the central portion of the organoids Ca and Cb is the brightness of the dark pixel in the peripheral portion. Replaced by a value. That is, the central portion of the organoid is darkly filled with the brightness value of the peripheral portion. On the other hand, in the organoids Cc and Cd in which the difference in brightness between the central portion and the peripheral portion is small, the change in the brightness value after the treatment is small. Further, around each organoid, the luminance value of the pixel in the background portion is replaced with the luminance value of the dark pixel in the peripheral portion of the organoid, so that the size of the organoid occupying the image becomes larger than the original size. That is, the organoid expands in the image.

Next, the maximum value filtering process is executed on the image data after the minimum value filtering process (step S104). The window size at this time is the same as the size in the minimum value filtering process. As a result, as shown in the image I2 shown in FIG. 3, the brightness value of the pixel in the peripheral portion of each organoid is replaced with the brightness value of the pixel in the background portion in the vicinity thereof. This causes the organoids expanded by the minimum filtering process to return to almost their original shape and size. On the other hand, the change in the luminance value is small in each pixel in the central part of the organoid.

That is, by executing the minimum value filtering process and the maximum value filtering process in this order, the inside of the living organoids Ca and Cb whose central part is brighter than the peripheral part while maintaining the shape and size of the organoids in the image. However, it will be filled with the brightness value of the pixels in the peripheral portion. Such a remarkable change does not occur in the dead organoids Cc and Cd in which the difference in brightness between the central portion and the peripheral portion is small.

Subsequently, the difference between the original image data and the image data after the maximum value filtering process is calculated (step S105). Specifically, the luminance value of the pixel at the corresponding position in the image I2 is subtracted from the luminance value of each pixel constituting the original image I0. In the difference image I3 corresponding to the image data obtained in this way, the central portion of the organoids Ca and Cb whose brightness value has been lowered by the above processing becomes high brightness, while the peripheral portion and the organoid of the organoids Ca and Cb whose brightness value changes little is small. The entire Cc and Cd and the background portion have low brightness. That is, by the processing up to this point, only the central portion of the living organoids Ca and Cb is extracted in the image.

In other words, the region in which the pixel having the luminance value equal to or higher than the predetermined value exists in the image I3 at this time is the region in which the living organoid exists in the original image I0. Such a region can be extracted by performing the binarization process using an appropriate threshold value (step S106). By the processing up to this point, living organoids can be identified from the organoids contained in the original image I0. The threshold value may be experimentally determined in advance, or may be dynamically determined by an appropriate threshold value determination method based on the brightness value of each pixel of the difference image. As the threshold value determination method, known techniques such as Otsu's method and Max Entropy Thresholding can be applied.

For example, if the purpose is to obtain the position or the number of living organoids in the original image I0, the purpose is achieved by the result of the processing so far. Further processing is performed to achieve the purpose, for example, to determine the area of individual organoids, or to compare live and dead organoids.

By extracting the central portion of the living organoids Ca and Cb in the process of difference calculation and binarization, the size of the organoids Ca and Cb in the binarized image is smaller than the original size. In order to restore this, an expansion process is executed on the binarized image (step S107). For example, in a binarized image, this purpose is achieved by expanding the high-luminance region outward by a predetermined number of pixels. This makes it possible to determine, for example, the area of living organoids.

The binarized image I5 represented by the image data thus created clearly separates the region occupied by the living organoids Ca and Cb from the other regions. By using this image I5 as a mask image for the original image I0, it is possible to perform various processing on the original image I0 based on the position information of the living organoid. An image obtained by subjecting the original image to various processing according to the purpose is created (step S108), and the result image is output as an output image (step S109). As an output mode, for example, display on a screen, recording on a recording medium, transmission to an external device via a communication line, or the like is possible.

As an example of the image processing result, the output image If shown in FIG. 3 shows the outline of the region extracted by applying the mask image I5 to the original image I0 with a thick and dark outline. By performing such image processing, it is possible to clearly distinguish between living and dead organoids Ca, Cb, Cc, and Cd contained in the image and present them to the user. The mode of image processing is not limited to this, and is arbitrary.

FIG. 4 is a diagram illustrating the effect of the processing of the present embodiment. As shown on the left side of FIG. 4A, an example of the luminance value profile obtained for each position on the broken line in the original image I0 is shown on the right side of the same figure. In the organoid Ca, which is bright as a whole, the brightness is lower at the peripheral portion of the organoid than the brightness of the background shown by the dotted line, while the brightness is sufficiently higher than that at the central portion. On the other hand, in the generally dark organoid Cb, the brightness in the central portion is higher than that in the peripheral portion, but the difference from the brightness in the background is small.

Thus, the image of a living organoid has the property that the central part is bright relative to the peripheral part, but for the absolute brightness value in the central part, other parts of the image, for example, dead. However, there are no distinctly different characteristics for organoids or backgrounds. Possible causes include individual variations in the thickness and color of organoids, variations in background brightness depending on the type and amount of medium, and variations in lighting conditions during imaging.

Therefore, even if an attempt is made to detect a living organoid by directly setting a threshold value with respect to the brightness value of the pixel of the original image, there is a risk of erroneous detection such as overlooking a relatively dark organoid. On the other hand, as shown in FIG. 4B, in the difference image I3 obtained in the processing process of the present embodiment, the region having relatively high brightness with respect to the surrounding pixels, that is, the living organoid Ca, While the central portion of Cb is emphasized, the pixels in the other regions have sufficiently low luminance values. Therefore, by setting the threshold Lth having a sufficient degree of separation with respect to the brightness value of the background portion shown by the dotted line on the right side of the figure, the probability of false detection is reduced and living organoids are reliably detected. It becomes possible.

FIG. 5 is a diagram illustrating the window size in the filtering process. As shown in FIG. 5A, the size Sw of the window W set in the vicinity of the central pixel Pc is about the same as the size Sc of the organoid (more generally, a cell or the like) C to be processed, for example. It is desirable to set the size to about 0.5 to 1 times the size Sc. When the cell or the like C is a substantially circular image in the image, the diameter of the circle can be regarded as the size Sc. With such a setting, it is possible to reliably fill the pixel in the central portion, which is relatively bright with respect to the peripheral portion in a living cell or the like, by the minimum value filter processing. As a result, when the difference between the processed image and the original image is obtained, it is possible to reliably extract the above-mentioned features such as living cells.

If the size Sw of the window W is too small with respect to the size Sc of the cells or the like C, the effect of this filling does not reach the center of the cells or the like C, and the difference in the brightness values when calculating the difference becomes small, resulting in good detection. It becomes difficult to do. Further, if the window size Sw is too large, there is a problem that the shape of cells or the like C in the image is greatly deformed by the processing, and the boundary between the peripheral portion of the cells or the like and the background becomes unclear. In the example of FIG. 5A, a neighborhood range of 5 × 5 pixels centered on the pixel Pc is set as the window W. As described above, in order to suppress the loss of information contained in the original image and obtain an appropriate filtering effect, the window size Sw is preferably about several pixels.

As shown in FIG. 5B, when the size Sc of cells or the like C is sufficiently larger than the pixel size Sp, setting the window size according to such a large cell or the like saves image information. It is not desirable from the viewpoint. In such a case, as shown in the figure, the above-mentioned filling is performed while suppressing the loss of image information by repeatedly executing the filter processing in which the window size Sw is set to several pixels. Is possible. Even in this case, it is desirable that the window size in the maximum value filtering process and the number of times the filtering process is executed are the same as those in the minimum value filtering process.

FIG. 6 is a diagram showing another example of image processing. In the above-mentioned processing, in order to restore the size of the region smaller than the original size during the binarization processing, expansion processing for a predetermined number of pixels is performed. Alternatively, for example, the following can be done. As shown in FIG. 6, the same is true up to the point where the original image I0 is subjected to various processing to obtain the binarized image I4. On the other hand, in the original image I0, since the peripheral portion of the organoid Ca or the like is darker than the background, it is possible to extract only the contour portion of the organoid Ca or the like by performing a known contour extraction process such as edge detection. ..

From such a contour image I6, information regarding the contour position of the organoid Ca or the like in the original image I0 can be obtained. By combining this information with the information indicating the region of the living organoid (described as "live region information" in FIG. 6) obtained from the binarized image I4, the position and contour of the living organoid can be obtained. An image I5a that can be used as a show mask image can be created. In addition, it is possible to individually determine the life or death of each organoid specified by the contour position information. As described above, the handling of the images after the difference image I3 is not limited to the above-mentioned ones, and various processing processes can be performed depending on the purpose.

FIG. 7 is a diagram showing another example of image processing. In the above processing example, as the final output image If, an image in which living organoids contained in the original image I0 are surrounded by a thick outline is created. The processing of the image for presenting the detection result is not limited to this, and various processes can be applied as described below. In the example shown in FIG. 7A, the output image Ifa is an image obtained by cutting out only the region indicated by the living organoids Ca and Cb from the original image I0. On the other hand, in the example shown in FIG. 7B, out of the organoids detected from the original image I0, only the regions occupied by the living organoids Ca and Cb, that is, the dead organoids Cc and Cd were cut out. The image is taken as the output image Ifb.

In addition, various image processing is possible, such as color-coding according to the life or death of the organoid, adding a marking for distinguishing between a living organoid and a dead organoid, and correcting the brightness according to the background brightness. In this way, if the region of the living organoid can be specified in the original image I0, it is possible to perform various image processing using the information and present it to the user.

As described above, in this embodiment, the characteristic difference is that the peripheral portion is dark and the central portion is bright in the living cells and the like in the image, whereas the peripheral portion and the central portion are dark in the dead cells and the like. To detect living cells and the like. Specifically, by performing the minimum value filtering process on the original image, the bright region in the central portion of the image of the living organoid is filled with the brightness value in the dark region in the peripheral portion. Then, by executing the maximum value filtering process, the size of the organoid expanded by the minimum value filtering process is restored.

By subtracting the image thus obtained from the original image, the central part of the filled living organoid can be extracted as a difference. That is, the region consisting of pixels having a brightness value equal to or higher than a predetermined value in the difference image can be distinguished from other portions as a region of living organoids. In this way, living organoids are detected in the image. For a variety of reasons, the absolute brightness of living organoids varies. In the present embodiment, since the region where the peripheral portion is dark and the central portion is relatively bright can be extracted, it is possible to suppress the occurrence of erroneous detection due to such variation and perform the detection with high accuracy.

The present invention is not limited to the above-described embodiment, and various modifications other than those described above can be made without departing from the spirit of the present invention. For example, in the above embodiment, organoids are mentioned as an example of "cells and the like", but the above method can be applied to various detection objects such as cells existing alone and spheroids.

Further, in the above embodiment, the living cell detection method of the present invention is applied to an original image captured in a bright field under visible light illumination, but the method for acquiring the original image is not limited to this. That is, if the image obtained by imaging shows the characteristic that "living cells and the like appear bright and dead cells and the like appear dark", images obtained by imaging methods other than the above are also applicable to the present invention. Can be. In particular, since the present invention can evaluate cells and the like in a non-staining and non-invasive manner, it is desirable to adopt a non-staining and non-invasive imaging method for imaging cells and the like.

Further, in the above embodiment, the present invention is used for the purpose of detecting where a living organoid is contained in the original image, but in addition to this, for example, a specific cell specified in the image. It is possible to use the present invention for determining whether or not a person is alive or dead. That is, by executing the above processing, it is possible to determine whether or not the cells or the like at the designated position are alive. In this case, the processing after the binarization processing is not always necessary.

Further, in the above embodiment, the living cell detection method according to the present invention is executed by a computer device having a general hardware configuration, but the present invention may be realized by a dedicated hardware device. Further, since the method of the present invention can be carried out by a general-purpose computer device, it can be used as a program for causing the computer device to perform the above processing, or as a computer-readable recording medium on which the program is recorded. In some embodiments, the present invention may be put into practical use.

As described above, as described by exemplifying a specific embodiment, in the present invention, for example, a region occupied by pixels having a brightness value equal to or higher than a threshold value in the difference is extracted, and the region is subjected to expansion processing of a predetermined size. The obtained region can be the region occupied by living cells. Since the detection method of the present invention extracts a prominent bright portion in the central part of a living cell or the like, the detected region tends to be smaller than that of an actual cell or the like. By executing the expansion treatment, it is possible to recover the size of such reduced cells and the like and detect a region close to the original size.

Further, for example, the size of the neighborhood range may be set according to the size of the detection target object in the original image. In the present invention, the minimum value filtering process is applied for the purpose of darkly painting the central portion of a living cell or the like, which is relatively bright with respect to the peripheral portion, with the brightness value of the peripheral portion. Therefore, the best result can be obtained by setting the window size in the filter processing, that is, the size of the vicinity range according to the size of the cell or the like which is the detection target.

More specifically, for example, the size of one side of the neighborhood range may be set to a value corresponding to the diameter of the detection object in the original image. With such a configuration, it is possible to reliably fill even the central part of a living cell or the like.

Further, for example, a step of accepting a setting input from the user regarding the size of the neighborhood range may be further provided. According to such a configuration, the user can adjust the window size in the filter processing, and the processing corresponding to the user's purpose can be realized.

Further, for example, the minimum value filtering process may be executed a plurality of times, and the maximum value filtering process may be executed the same number of times as the minimum value filtering process. In particular, when the size of the detection target is large relative to the pixel size, much of the information contained in the original image is lost and the shape of cells and the like becomes unclear if a wide neighborhood range is set according to the detection target. There is. Such a problem can be solved by executing multiple filtering processes in a relatively narrow neighborhood range.

Further, for example, the original image may be an image obtained by bright-field imaging of an unstained detection object. In such an image, the characteristic that the peripheral portion is dark and the central portion is bright appears prominently in living cells and the like. Therefore, by applying the living cell detection method according to the present invention, living cells and the like can be accurately detected. It becomes possible to detect.

The present invention detects living cells and the like from images of cells, cell masses and the like, and is particularly suitably applicable to the fields of medicine and biochemistry such as drug discovery screening.

C1-C4, Ca, Cb Living organoids C5-C10, Cc, Cd Dead organoids I0, Is Original image I1 Image after minimum value filtering I2 Image after maximum value filtering I3 Difference image I4 Binarized image I5, I5a Mask image If, Ifa, Ifb Output image Sw Window size

Claims (9)

The process of acquiring an original image including a visible light image of a cell or cell mass as a detection target, and
A step of executing a minimum value filter process based on the brightness value of the pixel and the neighboring pixels within a predetermined neighborhood range centered on the pixel for each pixel constituting the original image.
For each pixel constituting the image after the minimum value filtering process, the maximum value based on the brightness value of the pixel and the neighboring pixels in the neighborhood range having the same size as the neighborhood range in the minimum value filtering process centered on the pixel. The process of performing filtering and
The step of obtaining the difference between the original image and the image after the maximum value filtering process, and
A method for detecting a living cell, comprising a step of specifying a region occupied by a pixel having a brightness value equal to or higher than a predetermined threshold value as a region corresponding to the living cell in the difference. The raw according to claim 1, wherein a region occupied by pixels having a brightness value equal to or higher than the threshold value in the difference is extracted, and the region obtained by subjecting the region to an expansion treatment of a predetermined size is defined as a region occupied by the living cells. Cell detection method. The live cell detection method according to claim 1 or 2, wherein the size of the neighborhood range is set according to the size of the detection object in the original image. The live cell detection method according to claim 3, wherein the size of one side of the neighborhood range is set to a value corresponding to the diameter of the detection target in the original image. The live cell detection method according to any one of claims 1 to 4, further comprising a step of accepting a setting input from a user regarding the size of the neighborhood range. The living cell detection method according to any one of claims 1 to 5, wherein the minimum value filtering process is executed a plurality of times, and the maximum value filtering process is executed the same number of times as the minimum value filtering process. The live cell detection method according to any one of claims 1 to 6, wherein the original image is an image obtained by bright-field imaging of the unstained object to be detected. A minimum value filter based on the brightness value of the pixel and neighboring pixels within a predetermined neighborhood range centered on the pixel for each pixel constituting the original image including the visible light image of the cell or cell mass as the detection target. The process of executing the process and
For each pixel constituting the image after the minimum value filtering process, the maximum value based on the brightness value of the pixel and the neighboring pixels in the neighborhood range having the same size as the neighborhood range in the minimum value filtering process centered on the pixel. The process of performing filtering and
The step of obtaining the difference between the original image and the image after the maximum value filtering process, and
A program that causes a computer to execute a step of specifying a region occupied by pixels having a brightness value equal to or higher than a predetermined threshold value as a region corresponding to a living cell in the difference. A computer-readable recording medium on which the program according to claim 8 is recorded.

Images