Dyeing intensity acquisition method, program, and dyeing intensity acquisition device

Abstract

In a target region (42) corresponding to at least a part of a cell, first, a plurality of high-intensity pixels (422) having relatively high pixel staining intensities, which are staining intensities derived from pixel values, are identified. Next, a region staining intensity, which is the staining intensity of the target region (42), is acquired as an average value of the pixel staining intensities of the plurality of high-intensity pixels (422). As a result, it is possible to obtain an appropriate regional staining intensity in which false negative and false positive can be suppressed in staining determination.

Description

Methods, procedures, and devices for obtaining staining intensity Technical Field

[0001] This invention relates to a technique for obtaining the staining intensity of at least a portion of a region of each cell from a cell image representing a plurality of stained cells.

[0002] [Reference to relevant applications]

[0003] This invention claims the benefit of priority to Japanese Patent Application JP2023-152127, filed on September 20, 2023, the entire contents of which are incorporated herein by reference. Background Technology

[0004] In recent years, medical and biological research has been conducting analyses of single-cell units of biologically derived specimens, such as tissues and cells, with the aim of elucidating disease mechanisms, biological mechanisms, and drug action mechanisms. For example, in multiplex immunostaining analysis, various methods are used to stain biologically derived substances such as proteins, and quantitative and qualitative analyses are performed by measuring the staining status.

[0005] In Japanese Patent Application Publication No. 2022-123488 (Document 1), cell regions corresponding to cells, or regions of the central or peripheral parts of cells, are extracted from an image of stained tissue. The average concentration of the extracted region or the maximum concentration within the region is obtained as the "stain concentration". Then, when the stain concentration is above a predetermined threshold, the cell is determined to be "positive" relative to the stain; when the stain concentration is below the threshold, the cell is determined to be "negative" relative to the stain. "Positive" means, for example, that the cell has a considerable amount of a specific protein or that small intracellular organelles are present.

[0006] However, as shown in Reference 1, when determining whether a staining is positive or negative based on the average staining intensity (staining concentration in Reference 1) of the region extracted from the cell image, there is a concern that it may be classified as a false positive when the overall staining intensity is high enough, or as a false negative because the region with high staining intensity is small. Furthermore, when determining based solely on the maximum concentration within the extracted region, pixel values ​​are strongly influenced by anomalous pixels, resulting in insufficient accuracy. Summary of the Invention

[0007] The object of the present invention is to obtain an appropriate staining intensity of an object region corresponding to at least a portion of each cell from a cell image representing a plurality of stained cells.

[0008] The first aspect of the present invention is a staining intensity acquisition method, which acquires the staining intensity, i.e., the region staining intensity, of an object region corresponding to at least a portion of each cell from a cell image representing a plurality of stained cells; it comprises the following steps: a) in the object region, determining a plurality of high-intensity pixels whose staining intensity derived from pixel values ​​is relatively high; and b) acquiring the region staining intensity of the object region based on the pixel values ​​of the plurality of high-intensity pixels.

[0009] According to Method 1 of the present invention, an appropriate staining intensity can be obtained for the target area.

[0010] Method 2 of the present invention is the method for obtaining the staining intensity of Method 1, and the plurality of high-intensity pixels are a predetermined number of pixels with the highest staining intensity in the object region.

[0011] The third method of the present invention is the method for obtaining the staining intensity of the second method, and the predetermined number is proportional to the number of pixels contained in the object region.

[0012] The fourth aspect of the present invention is the dyeing intensity acquisition method of aspect 1 (or any one of aspects 1 to 3), and the plurality of high-intensity pixels are separated from each other by a predetermined distance or more.

[0013] The fifth aspect of the present invention is the method for obtaining the color intensity of aspect 1, and the above a) step includes the following steps: in the above object area, determining at least one peak pixel whose pixel color intensity is higher than that of the surrounding pixels; and determining the pixels in a circular area of ​​a predetermined size centered on the above at least one peak pixel as the above plurality of high intensity pixels.

[0014] Method 6 of the present invention is a staining intensity acquisition method of Method 1 (or any one of Methods 1 to 5), and the cell image described above represents a tissue specimen.

[0015] Method 7 of the present invention is any of the staining intensity acquisition methods of methods 1 to 6, and the object region is extracted from the cell image in the form of a region representing a whole cell.

[0016] The present invention, embodiment 8, is any of the staining intensity acquisition methods of embodiments 1 to 6, and further comprises the following steps before step a) above: step c) extracting a provisional target region corresponding to at least a portion of each of the cells from the cell image; and step d) dividing the provisional target region into an inner region and an outer region; wherein the target region is either the inner region or the outer region.

[0017] The present invention, embodiment 9, is a program that enables a computer to obtain the staining intensity, i.e., the region staining intensity, of an object region corresponding to at least a portion of each cell from a cell image representing a plurality of stained cells. The computer executes the program by performing the following steps: a) determining a plurality of high-intensity pixels in the object region whose staining intensity, i.e., pixel staining intensity, is relatively high based on pixel values; and b) obtaining the region staining intensity of the object region based on the pixel values ​​of the plurality of high-intensity pixels.

[0018] The present invention, embodiment 10, is a staining intensity acquisition device that acquires the staining intensity, i.e., the region staining intensity, of a target region corresponding to at least a portion of each cell from a cell image representing a plurality of stained cells. The device comprises: a pixel determination unit that determines, in the target region, a plurality of high-intensity pixels whose staining intensity, i.e., pixel staining intensity, derived from pixel values ​​is relatively high; and an intensity acquisition unit that acquires the region staining intensity of the target region based on the pixel values ​​of the plurality of high-intensity pixels.

[0019] The above-mentioned objects, as well as other objects, features, methods, and advantages, will become more apparent with reference to the accompanying drawings and through the following detailed description of the invention. Attached Figure Description

[0020] Figure 1 is a diagram showing the configuration of a computer that functions as a device for obtaining dyeing intensity.

[0021] Figure 2 is a block diagram showing the functional structure of the dyeing intensity acquisition device.

[0022] Figure 3 is a diagram showing the computer's operation flow.

[0023] Figure 4 shows a diagram representing a cell image.

[0024] Figure 5 is a diagram representing the object region.

[0025] Figure 6 is a diagram showing the location of high-intensity pixels determined in the object region.

[0026] Figure 7 shows a cell image displaying the determination results.

[0027] Figure 8 shows the state in which the entire object region is non-specifically stained.

[0028] Figure 9 is a diagram illustrating other operational examples of the dyeing intensity acquisition unit.

[0029] Figure 10 is a diagram illustrating other operational examples of the dyeing intensity acquisition unit.

[0030] Figure 11 is a flowchart illustrating further examples of other operations in the dyeing intensity acquisition section.

[0031] Figure 12 is a diagram illustrating further examples of other operations of the dyeing intensity acquisition section.

[0032] Figure 13 is a flowchart illustrating other action examples of the object area acquisition section.

[0033] Figure 14 is a diagram illustrating other action examples of the object area acquisition section.

[0034] Figure 15 is a diagram illustrating other action examples of the object area acquisition section. Detailed Implementation

[0035] Figure 1 is a diagram illustrating the configuration of a computer 5 that functions as a staining intensity acquisition device according to one embodiment of the present invention. The staining intensity acquisition device is a device that acquires the staining intensity (hereinafter referred to as "regional staining intensity" to distinguish it from the staining intensity of each pixel as described later) of a target region corresponding to at least a portion of each cell from a cell image. The cell image is a digital image representing multiple cells stained with a biologically derived substance such as a specific protein or RNA (ribonucleic acid), for example, an image obtained by photographing a biologically derived specimen obtained by immunostaining. The cell image may also be one of multiple images obtained each time in a multiple immunostaining process involving repeated staining and destaining.

[0036] Computer 5 has the general structure of a computer system, including a CPU 51, GPU 52, ROM 53, RAM 54, hard disk 55, display 56, input unit 57, reading device 58, communication unit 59, and bus 50. The CPU 51 performs various arithmetic operations. The GPU 52 performs various arithmetic operations related to image processing at high speed. The ROM 53 stores the basic program. The RAM 54 stores various information. The hard disk 55 stores information. The display 56 is a display unit for displaying images and other information.

[0037] The input unit 57 includes a keyboard 57a and a mouse 57b for accepting input from the operator. The reading device 58 reads information from a computer-readable recording medium 9 such as an optical disc, magnetic disk, magneto-optical disk, or memory card. The display 56, keyboard 57a, mouse 57b, and reading device 58 are connected to the bus 50 via an interface (I / F). The communication unit 59 transmits and receives signals with devices external to the computer 5.

[0038] Bus 50 is a signal loop connecting CPU 51, GPU 52, ROM 53, RAM 54, fixed disk 55, display 56, input unit 57, reading device 58, and communication unit 59. An external imaging device 8 is connected to bus 50 via I / F. The imaging device 8 acquires images of the specimen as cell images and stores the data on the fixed disk 55.

[0039] In computer 5, program 91 is read from recording medium 9 beforehand via reading device 58 and stored on fixed disk 55. Program 91 may also be stored on fixed disk 55 via network. CPU 51 and GPU 52 perform arithmetic processing according to program 91 while utilizing RAM 54 or fixed disk 55. CPU 51 and GPU 52 function as arithmetic units in staining intensity acquisition device 1. In addition to CPU 51 and GPU 52, other configurations that function as arithmetic units may also be used.

[0040] Figure 2 is a block diagram showing the functional configuration of the staining intensity acquisition device 1, which is implemented by the computer 5 performing calculations and processing according to program 91. Figure 2 also shows configurations other than the functions of the staining intensity acquisition device 1. In Figure 2, the staining intensity acquisition unit 13 represents the main functional configuration of the staining intensity acquisition device 1. The storage unit 11 and the target area acquisition unit 12 may also be included in the functional configuration of the staining intensity acquisition device 1. The computer 5 further implements the functional configuration of the staining determination unit 14, thereby making the computer 5 a device for determining whether each cell in a cell image is positive or negative for staining.

[0041] All or part of each functional component can also be implemented using dedicated circuitry. Furthermore, these functional components can be implemented using multiple computers. In the functional configuration shown in Figure 2, the storage unit 11 is mainly implemented using RAM 54 and a fixed disk 55. The object area acquisition unit 12, the staining intensity acquisition unit 13, the staining determination unit 14, and the display control unit 15 are implemented using a CPU 51, a GPU 52, a ROM 53, RAM 54, a fixed disk 55, and their peripherals.

[0042] The storage unit 11 stores images of biological specimens (hereinafter also simply referred to as "specimens") acquired by the imaging device 8, namely cell images. The stored images are more precisely "image data 81" as shown in FIG2. In the following description, the processing of the images is more precisely the processing of image data 81.

[0043] Figure 3 is a diagram illustrating the action flow of computer 5, which shows the action of obtaining the staining intensity of each object region in the cell image, i.e., the region staining intensity (which can also be regarded as the staining intensity of each cell), and determining whether the object region is positive or negative relative to the staining.

[0044] The specimen is first photographed by the imaging device 8 controlled by the computer 5, or by the operation of the imaging device 8, and cell images are prepared in the storage unit 11 (step S11). As illustrated in FIG4, the cell image 4 stored in the storage unit 11 represents a plurality of cells 41. In FIG4, each cell 41 is simplified and represented by a circle.

[0045] Next, the target region acquisition unit 12 acquires a region from the cell image 4 that corresponds to at least a portion of each cell 41, as the target region (step S12). When the processing object, i.e., the cell image 4, is one of multiple images obtained from each staining in a multiple immunostaining method involving repeated staining and destaining, a region corresponding to the target region is acquired from a cell image generated using a staining different from that of the cell image 4 (i.e., different staining images of the same specimen), and this region is used as the target region of the cell image 4. Furthermore, the function of the region segmentation unit 121 within the target region acquisition unit 12 will be described later. Hereinafter, the multiple target regions acquired by the target region acquisition unit 12 will be referred to as a "target region group".

[0046] A target region can be a so-called "cellular region" corresponding to a single cell 41, or it can be a part of a cellular region, such as the region corresponding to the cell nucleus. Various known methods can be used to obtain the target region. For example, various known methods for obtaining cellular regions, various known methods for obtaining nuclear regions, and known methods for obtaining at least a portion of a cellular region that is not bound by tissue can be used. When performing single-cell analysis, the target region and the cell preferably correspond one-to-one. That is, the target region is at least a portion of the cellular region, and the cellular region and the target region correspond one-to-one. Thus, the target region is the region corresponding to at least a portion of each cell. If the computer 5 has the function of extracting the target region, it may not necessarily have the function of extracting the cellular region.

[0047] Next, the staining intensity acquisition unit 13 selects the object region as the initial treatment target (step S13). Figure 5 is a diagram illustrating the object region 42 when the object region 42 is a cellular region. In the object region 42, there is a stained region 421. In Figure 5, for ease of understanding, although the region 421 is represented by a circle with parallel diagonal lines, in reality, the region 421 has varying shades, various shapes, and sometimes unclear outlines.

[0048] The staining intensity acquisition unit 13 includes a pixel determination unit 131 and an intensity acquisition unit 132. The pixel determination unit 131 determines, within the target region 42, a plurality of pixels (hereinafter referred to as "high-intensity pixels") with relatively high staining intensity (hereinafter referred to as "pixel staining intensity") derived from the pixel value (step S14). The term "staining intensity" in "region staining intensity" and "pixel staining intensity" refers to the intensity of staining as seen in a cell image. For example, in an image obtained by capturing the color development produced by staining using a bright-field imaging system, if a strongly stained area appears darker, the darker the area, the higher the staining intensity. Conversely, in an image obtained by capturing the color development produced by staining with fluorescent dyes, if a strongly stained area appears brighter, the brighter the area, the higher the staining intensity.

[0049] Therefore, when the cell image is a monochrome image, there are cases where a larger pixel value results in a higher pixel staining intensity, and conversely, a larger pixel value results in a lower pixel staining intensity. When the cell image is a color image, there are cases where a larger value of a specific color element of a pixel, or a larger value obtained through calculation from multiple color elements, results in a higher pixel staining intensity, and conversely, a larger value results in a lower pixel staining intensity. It should be noted that when the cell image is a color image, color decomposition or color extraction can be performed beforehand to extract only preset staining hue components, and the resulting image can be used as a new cell image for processing. When the cell image is a monochrome image, the image after image processing can also be used as a new cell image for processing. As will be described later, the relationship between the intensity of regional staining derived from the pixel values ​​of multiple high-intensity pixels and the brightness or hue of the object region 42 within the image is also the same.

[0050] Generally speaking, "pixel chromatic intensity" is a value derived from the pixel value of a single pixel or from the pixel values ​​of multiple pixels contained in a region (pixel chromatic intensity can also be the pixel value itself), representing the intensity of chromaticity. More precisely, "pixel chromatic intensity" is the value obtained from the pixel value of a single pixel, while "region chromatic intensity" is the value obtained from the pixel values ​​of multiple high-intensity pixels contained in the object region 42.

[0051] As described above, if the pixel tint intensity increases, the pixel value (which includes the value of one color component of the color pixel value or the value derived from multiple color components) simply increases or simply decreases. Therefore, the pixel determination unit 131 of the tint intensity acquisition unit 13 does not actually calculate the pixel tint intensity, but can determine a number of high-intensity pixels with relatively high pixel tint intensity in the target area 42. For example, the pixel determination unit 131 determines a predetermined number of pixels from the largest pixel value or a predetermined number of pixels from the smallest pixel value as high-intensity pixels. Thus, a predetermined number of pixels with the highest pixel tint intensity can be easily determined as a number of high-intensity pixels in the target area 42. It should be noted that the so-called "predetermined number of pixels with the highest pixel tint intensity" refers to the predetermined number of pixels located at the top when the pixels are arranged in descending order of pixel tint intensity.

[0052] Figure 6 illustrates the positions of the four high-intensity pixels 422 identified in the object region 42, marked with "+". While the number of high-intensity pixels 422 identified in each object region 42 is not particularly limited, in a preferred embodiment, the number of high-intensity pixels 422 is proportional to the number of pixels contained in each object region 42 in one image. In other preferred embodiments, the number of high-intensity pixels 422 is a fixed value independent of the size of the object region 42. The ratio of the number of high-intensity pixels 422 to the number of pixels in the object region 42, and the number of high-intensity pixels 422, can be appropriately set by the actual size of the object region 42, the resolution of the cell image, and the characteristics of the staining (e.g., the extent to which proteins or intracellular organelles are present as the object, the size of the object, and the ease of staining).

[0053] The ratio of the number of high-intensity pixels 422 to the number of pixels in the object region 42 can also be automatically determined through learning. For example, for cell images in various staining conditions, the computer learner can learn the differences between positive and negative positivity in multiple object regions 42 and the optimal combination of the number of high-intensity pixels 422. When the input image is input to the learner, the ratio of the number of high-intensity pixels 422 to the number of pixels in the object region 42 is output.

[0054] Next, the intensity acquisition unit 132 acquires the average value of the pixel color intensity of the determined plurality of high-intensity pixels 422 as the region color intensity of the target region 42 (step S15). As described above, pixel values ​​can be used instead of pixel color intensity, so the average value of the pixel values ​​of the high-intensity pixels 422 can be acquired as the region color intensity in step S15. Of course, the pixel color intensity in each high-intensity pixel 422 can also be calculated, and the average value of the pixel color intensity can be acquired as the region color intensity of the target region 42. In this way, the intensity acquisition unit 132 acquires the region color intensity of the target region 42 based on the pixel values ​​of the plurality of high-intensity pixels 422. When the number of high-intensity pixels 422 is proportional to the number of pixels contained in the target region 42, the region color intensity can be acquired while suppressing the influence of the size of the target region 42.

[0055] When the staining intensity is obtained for one object region 42, the staining intensity acquisition unit 13 checks whether there is a next object region 42 within the object region group for which the staining intensity has not been determined (step S16). If so, the next object region 42 is selected and the staining intensity is obtained (steps S13 to S15). When the staining intensity is obtained for all object regions 42 (step S16), the process is transferred to the staining determination unit 14.

[0056] In the staining determination unit 14, a threshold is preset by the operator via the input unit 57, and the staining intensity of each object region 42 in the object region group is compared with the threshold. When the staining intensity of an object region 42 is higher than the threshold, the object region 42 is determined to be positive for staining (step S17). Furthermore, the smaller the average pixel value of the high-intensity pixels contained in the object region 42, the higher the staining intensity. When the staining determination unit 14 directly compares this average pixel value with the threshold, if the average pixel value is lower than the threshold, the object region 42 is determined to be positive for staining. The determination result 82 for each object region 42 is stored in the storage unit 11.

[0057] The determination result 82 is displayed on the display 56 under the control of the display control unit 15. For example, as shown in FIG7, the positive object region 42a in the object region 42 of the cell image 4 is colored, and the cell image 4 is displayed on the display 56.

[0058] In the staining intensity acquisition device 1, staining determination, i.e., determination of whether staining is positive, can be appropriately performed by utilizing the pixel staining intensity of high-intensity pixels 422. For example, as shown in FIG8, sometimes the entire object region 42 is non-specifically stained, and the entire object region 42 reaches a maximum staining intensity of about 60%. Although it should not be determined as positive under such staining conditions, as in the past, when the average pixel staining intensity of the entire object region 42 is obtained as the region staining intensity of the object region 42, the object region 42 will be falsely judged as positive when the threshold is 50%. On the other hand, although it should be judged as positive in the case of FIG6, because there are few areas with high staining intensity, even if the average pixel staining intensity of the entire object region 42 is used as the region staining intensity, if the value is 40%, the object region 42 will be falsely judged as negative.

[0059] In contrast, when using the pixel staining intensity of high-intensity pixels 422, if the average pixel staining intensity of high-intensity pixels 422 is obtained as the region staining intensity, then the obtained region staining intensity is, for example, 80%, and the object region 42 will be determined as positive. Thus, by utilizing the pixel staining intensity of high-intensity pixels 422 (i.e., using pixel values), in the case of regions with strong staining, it is possible to appropriately determine them as positive, and false positives and false negatives can be suppressed.

[0060] The staining determination using high-intensity pixels 422 is particularly suitable for performing staining determination on cell images that have been stained in a way that results in dense nuclear or cytoplasmic staining of only a portion of the target region 42. Furthermore, appropriate staining determination can still be performed even when noise, known as background noise, is present, such as when a portion or the entire cell tissue is lightly and nonspecifically stained. Background noise is more likely to occur in tissue specimen staining than in cultured cells; therefore, staining determination using high-intensity pixels 422 is suitable for staining tissue specimens.

[0061] Furthermore, staining determination using high-intensity pixels 422 is suitable for staining proteins with high localization or uneven distribution within cells, even in the presence of background values. For example, it is more suitable for staining fibrillarin, which is locally present in the nucleolus, than for staining actin, which forms the cytoskeleton. It is also suitable for staining images with a low signal-to-noise ratio (S / N). For instance, it is more suitable for staining determination using enzyme-antibody methods with chromogenic agents and bright-field optical systems than for staining determination using fluorescent dyes and fluorescent optical systems.

[0062] It should be noted that the same effect of using the high-intensity pixel 422 to determine the regional color intensity applies to the case described later, where peak pixels are used to determine high-intensity pixels to determine the regional color intensity.

[0063] Figures 9 and 10 are diagrams illustrating other operational examples of the staining intensity acquisition unit 13. Figure 9 shows an example of an object region 42 that should be determined as negative. In the object region 42 of Figure 9, there is one strongly stained region 421a and several small weakly stained regions 421b that should not be considered stained. In the example of Figure 9, when a high-intensity pixel 422 is identified, if a predetermined number of pixels with high staining intensity are extracted from region 421a, the average staining intensity of the high-intensity pixel 422 becomes higher, and the object region 42 is mistakenly determined as positive.

[0064] Therefore, in the improved color intensity acquisition unit 13, a plurality of high-intensity pixels 422 are acquired in the form of pixels that are more than or equal to each other at a predetermined distance. Specifically, as shown in FIG10, firstly, the pixel 422a with the highest pixel color intensity is determined as a high-intensity pixel 422. Next, among the pixels that are more than or equal to the radius of the circle of the dashed line 43 from the pixel 422a, the pixel 422b with the highest pixel color intensity is determined as the next high-intensity pixel 422. Next, among the pixels that are more than or equal to the radius of the circle of the dashed line 43 from any of the determined high-intensity pixels 422a and 422b, the pixel 422c with the highest pixel color intensity is determined as the next high-intensity pixel 422. The above operation is repeated to determine a predetermined number of high-intensity pixels 422 (pixels 422a to 422d shown in FIG10).

[0065] Since the pixel staining intensities of pixels 422b, 422c, and 422d are not high, the average pixel staining intensity of the high-intensity pixel 422, i.e., the region staining intensity, becomes low. As a result, the object region 42 can be correctly identified as negative. By separating multiple high-intensity pixels 422 from each other, the region staining intensity can be determined while suppressing the influence of the determined high-intensity region.

[0066] The distance at which the multiple high-intensity pixels 422 should be separated, i.e., the radius of the circle shown by the dashed line 43 in Figure 10, can be automatically determined. For example, this distance can be determined as a value that is proportional to the average radius when the multiple cell regions are considered as circles and inversely proportional to the square root of the number of high-intensity pixels 422 to be determined. Thus, the "predetermined distance" at which the multiple high-intensity pixels 422 should be separated can be different for each object region 42.

[0067] Next, further examples of other operations of the staining intensity acquisition unit 13 in FIG2 will be described with reference to FIGS. 11 and 12. In this example, in step S14 of FIG. 3, the pixel determination unit 131 performs the operation shown in FIG. 11.

[0068] First, as shown in FIG12, the pixel determination unit 131 determines at least one peak pixel 423 in the target region 42 (step S141). The peak pixel 423 is a pixel in the target region 42 whose pixel coloring intensity is higher than that of the surrounding pixels; that is, a pixel whose pixel coloring intensity becomes a peak when viewed in two dimensions. However, in order to exclude peak pixels in regions with low coloring intensity, a lower limit is set for the pixel coloring intensity used to determine the peak pixel 423. Furthermore, even if a pixel exists at a peak position, if the number of surrounding pixels forming the peak is small, the pixel can be excluded from the peak pixels. A peak pixel refers to a pixel with a pixel coloring intensity of a predetermined intensity or higher and located at the peak vertex of a predetermined size or higher (coloring intensity).

[0069] Next, the pixel determination unit 131 determines the pixels within a predetermined-size circular region 424 centered on at least one peak pixel 423 as a plurality of high-intensity pixels (step S142). The radius of the circular region 424 may be all the same, or the radius may be larger as the pixel coloring intensity of the peak pixel 423 increases. That is, the "predetermined size" of the circular regions 424 may be all the same, or it may be different for each circular region 424.

[0070] Subsequently, the intensity acquisition unit 132 acquires the average value of the pixel color intensity of all pixels within at least one circular region 424, and uses this average value as the region color intensity of the target region 42. When the number of circular regions 424 is two or more, after calculating the average value of the pixel color intensity of each circular region 424, the average value of the multiple average values ​​can be further averaged to obtain the region color intensity of the target region 42. In this way, the intensity acquisition unit 132 acquires the region color intensity of the target region 42 based on the pixel color intensity of the high-intensity pixels within a circular region 424 of a predetermined size centered on at least one peak pixel 423.

[0071] The determination of peak pixel 423 can be easily implemented, for example, using the open-source computer vision library "OpenCV". By using peak pixel 423 to obtain the regional staining intensity of object region 42, a more accurate staining determination can be made when staining such as multiple positive spots sporadically exist within object region 42. Therefore, obtaining the regional staining intensity using peak pixel 423 is suitable for cases where multiple peak pixels 423 are determined for one object region 42 in pixel determination unit 131. Staining determination using peak pixel 423 is particularly suitable for performing staining determination on cell images that have undergone staining where only a portion of object region 42 is densely stained with nuclear or cytoplasmic staining.

[0072] Next, other operation examples of the object region acquisition unit 12 in FIG2 will be described with reference to FIGS. 13 to 15. In other operation examples, the object region acquisition unit 12 performs steps S21 and S22 of FIG. 13 instead of step S12 of FIG. 3. The object region acquisition unit 12 first extracts a set of provisional object regions from the cell image 4 (step S21). A provisional object region may be a so-called "cell region" corresponding to a whole cell, or it may be a part of a cell region, such as the region corresponding to the cell nucleus. When performing single-cell analysis, it is preferable that the provisional object region corresponds to the cell in a one-to-one manner. That is, the provisional object region is at least a part of the cell region, and the cell region corresponds to the provisional object region in a one-to-one manner. In this way, the provisional object region is the region corresponding to at least a part of each cell. If the computer 5 has the function of extracting provisional object regions, it may not be necessary to have the function of extracting cell regions.

[0073] Figure 14 illustrates a provisional target area 44. Although the area 421 represented by the circle with parallel diagonal lines indicates the area to be colored, it actually has various sizes and shapes, and the area 421 also has varying shades. Furthermore, the outline of the area 421 may be unclear. The target area acquisition unit 12 has the function of the area division unit 121 shown in Figure 2, which divides each provisional target area 44 into an inner inner area 441 and an outer peripheral area 442 (step S22). The outer peripheral area 442 is an area of ​​fixed width along the outer peripheral edge of the provisional target area 44. The inner area 441 is an area further inward than the outer peripheral area 442. Next, the target area acquisition unit 12 acquires either the inner area 441 or the outer peripheral area 442 as the target area. That is, it acquires the inner area 441 of all provisional target areas 44 as the target area, or it acquires the outer peripheral area 442 of all provisional target areas 44 as the target area. Next, execute steps S13 to S17 as shown in Figure 3.

[0074] That is, when the inner region 441 is taken as the target region, using step S14 in FIG3, pixels with relatively high pixel color intensity among the pixels in the inner region 441 are identified as high-intensity pixels 422. In FIG14, high-intensity pixels 422 are illustrated by a "+" symbol. On the other hand, when the outer peripheral region 442 is taken as the target region, as illustrated in FIG15, using step S14, pixels with high pixel color intensity among the pixels in the outer peripheral region 442 are identified as high-intensity pixels 422. Next, the region color intensity of the target region is obtained (step S15). Furthermore, after obtaining the region color intensity of all target regions (step S16), color determination is performed on all target regions (step S17).

[0075] By setting the target region as the inner region 441, staining determination of the inner region 441 can still be appropriately performed even when non-specific staining occurs in the outer region 442. For example, when staining and detecting specific proteins in or near the cell nucleus, the influence of cell membrane staining can be excluded. Furthermore, the influence of staining from adjacent cells can also be excluded. When the target region is set as the outer region 442, staining determination of the outer region 442 can still be appropriately performed even when non-specific staining occurs in the inner region 441. For example, when staining and detecting specific proteins in the cell membrane, the influence of staining outside the cell membrane can be excluded.

[0076] Of course, the provisional target region 44 is not limited to the cellular region (the region representing the entire cell). For example, the provisional target region 44 can also be extracted in the form of the nucleus and its surrounding area, with the region inside the nucleus designated as the internal region 441 and the region on the periphery of the nucleus designated as the peripheral region 442. By dividing the provisional target region 44 into the internal region 441 and the peripheral region 442 and designating only one as the target region, artifacts in the analysis can be effectively suppressed, and more robust regional staining intensity can be obtained.

[0077] Various modifications can be made to the above implementation methods.

[0078] In the example shown in Figure 6, a predetermined number of pixels with the highest pixel color intensity are identified in the target region 42 as a plurality of high-intensity pixels 422. In the example shown in Figure 12, at least one peak pixel 423 and its surrounding pixels are identified as a plurality of high-intensity pixels. Thus, high-intensity pixels can be determined by various methods as long as they are pixels with relatively high pixel color intensity. Preferably, the high-intensity pixels are included within the upper 50% (more preferably within the upper 30%) of all pixels in the target region 42 where pixel color intensity is higher. That is, high-intensity pixels can be selected from the upper 50% of pixels with higher pixel color intensity using various methods. Then, the regional color intensity of the target region 42 is obtained by applying various operations to the pixel values ​​of the plurality of high-intensity pixels.

[0079] The action flow shown in Figure 3 can be modified in various ways within the same scope. For example, it can be an action that, each time the color intensity of an object region 42 is obtained, compares the color intensity of that region with a threshold and moves on to obtaining the color intensity of the next object region 42.

[0080] The method using either the inner region 441 or the outer region 442 shown in Figures 13 to 15 can also be applied to the method shown in Figures 11 and 12 for obtaining the regional color intensity using the peak pixel 423.

[0081] The staining method in this embodiment is not limited to multiple staining or staining using antibodies. This embodiment can be used for staining determination in various staining processes, as long as the substance is of biological origin. This embodiment is preferably applicable to situations where staining reveals varying degrees of intensity in cells. Furthermore, it can also be applied to staining specific proteins.

[0082] The components in the above-described embodiments and their variations can be appropriately combined as long as they do not contradict each other.

[0083] Although the invention has been described and illustrated in detail above, the foregoing description is illustrative only and not limiting. Therefore, it can be said that various modifications or methods can be implemented without departing from the scope of the invention.

[0084] Explanation of reference numerals in the attached figures

[0085] 1: Dyeing intensity acquisition device

[0086] 4: Cell images

[0087] 5: Computer

[0088] 13: Obtaining Staining Intensity

[0089] 41: Cell

[0090] 42: Object Region

[0091] 44: Provisional Object Area

[0092] 91: Program

[0093] 131: Pixel Determination Unit

[0094] 132: Strength Acquisition Section

[0095] 422: High-intensity pixels

[0096] 423: Peak pixels

[0097] 424: (Circular area centered on the peak pixel)

[0098] 441: Internal Area

[0099] 442: Peripheral area

[0100] S14, S15, S21, S22, S141, S142: Steps

Claims

1. A method for obtaining staining intensity, comprising obtaining staining intensity, i.e., region staining intensity, of an object region corresponding to at least a portion of each cell from a cell image representing a plurality of stained cells; comprising the steps of: a) determining, in the object region, a plurality of high-intensity pixels whose staining intensity derived from pixel values ​​is relatively high; and b) obtaining the region staining intensity of the object region based on the pixel values ​​of the plurality of high-intensity pixels.

2. The method for obtaining staining intensity according to claim 1, wherein, The plurality of high-intensity pixels are a predetermined number of pixels with the highest pixel coloring intensity in the object region.

3. The method for obtaining staining intensity according to claim 2, wherein, The predetermined number is proportional to the number of pixels contained in the object region.

4. The method for obtaining staining intensity according to claim 1, wherein, The plurality of high-intensity pixels are separated from each other by a predetermined distance or more.

5. The method for obtaining staining intensity according to claim 1, wherein, Step a) comprises the following steps: in the object region, determining at least one peak pixel whose pixel staining intensity is higher than that of the surrounding pixels; and determining pixels in a circular region of a predetermined size centered on the at least one peak pixel as the plurality of high-intensity pixels.

6. The method for obtaining staining intensity according to claim 1, wherein, The cell images represent tissue specimens.

7. The method for obtaining staining intensity according to any one of claims 1 to 6, wherein, The object region is extracted from the cell image in the form of a region representing a single cell as a whole.

8. The method for obtaining staining intensity according to any one of claims 1 to 6, wherein, Prior to step a), the following steps are further included: c) extracting a provisional object region from the cell image corresponding to at least a portion of each cell; and d) dividing the provisional object region into an inner region and a peripheral region; wherein the object region is either the inner region or the peripheral region.

9. A program that causes a computer to obtain, from a cell image representing a plurality of stained cells, the staining intensity, i.e., region staining intensity, of an object region corresponding to at least a portion of each cell, wherein the computer executes the program to cause the computer to perform the following steps: a) determining, in the object region, a plurality of high-intensity pixels with relatively high staining intensities derived from pixel values, i.e., pixel staining intensities; and b) obtaining the region staining intensity of the object region based on the pixel values ​​of the plurality of high-intensity pixels.

10. A staining intensity acquisition apparatus for acquiring staining intensity, i.e., region staining intensity, of a target region corresponding to at least a portion of each cell from a cell image representing a plurality of stained cells; comprising: a pixel determination unit that determines a plurality of high-intensity pixels in the target region whose staining intensity, i.e., pixel staining intensity, is relatively high based on pixel values; and an intensity acquisition unit that acquires the region staining intensity of the target region based on the pixel values ​​of the plurality of high-intensity pixels.

Images