Information processing device, information processing method, and information processing program

Abstract

This invention provides an information processing device, an information processing method, and an information processing program that can objectively evaluate structural changes in decellularized tissue. [Solution] The information processing device 10 includes a processor, which acquires a first image, which is an image of biological tissue, generates a pseudo-image by applying predetermined image processing to the first image, constructs a structure map showing the distribution of feature quantities of the first image and the pseudo-image, acquires a second image, which is an image of decellularized tissue from which cellular components have been removed from the biological tissue, maps the feature quantities of the second image to the structure map, and calculates a distance scale between the feature quantities of the second image in the structure map and predetermined reference points in the structure map.

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and an information processing program.

Background Art

[0002] In recent years, decellularized tissues composed of extracellular matrix (ECM) from which cell components have been removed from living tissues of animals such as humans, pigs, and cows have attracted attention. Decellularized tissues are being put into practical use as biological substitute materials for transplantation and regenerative medicine for various living tissues such as hard tissues such as bone, and soft tissues such as heart valves, blood vessels, and skin.

[0003] Specifically, research has been conducted on decellularization using various methods such as chemical methods and physical methods. At present, changes in the structure of ECM are inevitable regardless of the method used. Therefore, optimization of the decellularization process is being attempted according to the properties of the decellularized tissue's source and transplantation site (such as tissue density, cell density, fat content, and structure), as well as the characteristics and uses required for the final product.

[0004] In Non-Patent Document 1, the following three are proposed as evaluation criteria for decellularized tissues. The first criterion relates to the amount of residual DNA. Specifically, it is preferably less than 50 ng of DNA per 1 mg in dry weight. The second criterion relates to the length of residual DNA strands. Specifically, it is preferably less than 200 bp. The third criterion relates to the degree of residual cells. Specifically, it is preferably that cell residues are removed when observing the stained decellularized tissue.

[0005] Furthermore, it has become clear that evaluating the structural changes of the extracellular membrane (ECM) before and after decellularization is important for optimizing the decellularization process. For example, Non-Patent Literature 2 discloses evaluating the suitability of decellularized porcine aorta as vascular graft material by measuring its protein permeability. Here, it is stated that the degree of structural change of the ECM differs depending on the decellularization method, and that protein permeability reflects these differences in ECM structure. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Peter M. Crapo, Thomas W. Gilbert, Stephen F. Badylak, "An overview of tissue and whole organ decellularization processes," Biomaterials, 2011, 32(12), 3233-3243. [Non-Patent Document 2] Pingli Wu, Tsuyoshi Kimura, Hiroko Tadokoro, Kwangwoo Nam, Toshiya Fujisato, Akio Kishida, "Relation between the tissue structure and protein permeability of decellularized porcine aorta," Materials Science and Engineering: C, 2014, 43, 465-471. [Overview of the Initiative] [Problems that the invention aims to solve]

[0007] Although methods such as Non-Patent Document 2 have been proposed for evaluating changes in ECM structure before and after decellularization, a universally applicable scale has not been established, and evaluations are still often based on the subjective judgment of the evaluator. Therefore, interpreting and comparing various decellularization methods regarding changes in ECM structure has been difficult, hindering optimization.

[0008] This disclosure provides an information processing device, an information processing method, and an information processing program that can objectively evaluate structural changes in decellularized tissue. [Means for solving the problem]

[0009] A first aspect of this disclosure is an information processing device comprising a processor, the processor acquires a first image which is an image of biological tissue, generates a pseudo-image by applying predetermined image processing to the first image, constructs a structure map showing the distribution of feature quantities of the first image and the pseudo-image, acquires a second image which is an image of decellularized tissue from which cellular components have been removed from the biological tissue, maps the feature quantities of the second image to the structure map, and calculates a distance scale between the feature quantities of the second image in the structure map and predetermined reference points in the structure map.

[0010] A second aspect of this disclosure is an information processing method, wherein a computer performs the following processes: acquires a first image which is an image of biological tissue; generates a pseudo-image by applying predetermined image processing to the first image; constructs a structure map showing the distribution of feature quantities of the first image and the pseudo-image, acquires a second image which is an image of decellularized tissue from which cellular components have been removed from the biological tissue; maps the feature quantities of the second image to the structure map; and calculates a distance scale between the feature quantities of the second image in the structure map and predetermined reference points in the structure map.

[0011] A third aspect of this disclosure is an information processing program which causes a computer to perform the following processes: acquire a first image which is an image of biological tissue; generate a pseudo-image by applying predetermined image processing to the first image; construct a structure map showing the distribution of feature quantities of the first image and the pseudo-image, acquire a second image which is an image of decellularized tissue from which cellular components have been removed from the biological tissue; map the feature quantities of the second image to the structure map; and calculate the distance scale between the feature quantities of the second image in the structure map and predetermined reference points in the structure map. [Effects of the Invention]

[0012] According to the above embodiments, the information processing device, information processing method, and information processing program of the present disclosure can objectively evaluate structural changes in decellularized tissue. [Brief explanation of the drawing]

[0013] [Figure 1] This is a diagram illustrating the schematic configuration of an information processing system. [Figure 2] This figure shows an example of a biological tissue image. [Figure 3] This figure shows an example of a decellularized tissue image. [Figure 4] This is a block diagram showing an example of the hardware configuration of an information processing device. [Figure 5] This figure shows an example of a pseudo-image. [Figure 6] This is a diagram showing an example of a structure map. [Figure 7] This figure shows an example of a distance scale. [Figure 8] This figure shows an example of the information that will be output. [Figure 9] This is a diagram to explain the correction process. [Figure 10] This is a diagram to explain the correction process. [Figure 11] This is a flowchart illustrating an example of information processing. [Figure 12] This figure shows the structural map related to Example 1. [Figure 13]It is a diagram showing a structure map according to Example 2.

Embodiments for Carrying Out the Invention

[0014] Hereinafter, an example of an embodiment of the disclosed technology will be described while referring to the drawings. In each drawing, the same or equivalent components and parts are given the same reference numerals, and duplicate descriptions are omitted. Also, the dimensional ratios in the drawings are exaggerated for the convenience of explanation and may be different from the actual ratios.

[0015] Referring to FIGS. 1 to 3, the information processing system 100 according to the present embodiment will be described. The information processing system 100 is a system for evaluating decellularized tissue. The decellularized tissue is composed of an extracellular matrix (ECM: Extracellular Matrix) obtained by removing cell components from living tissues of animals such as humans, pigs, and cows. The decellularized tissue is being put into practical use as a biological substitute material for transplantation and regenerative medicine for various living tissues such as hard tissues such as bone and soft tissues such as heart valves, blood vessels, and skin.

[0016] FIG. 1 is a schematic diagram showing an example of the configuration of the information processing system 100. As shown in FIG. 1, the information processing system 100 includes an information processing device 10 and an imaging device 12. Each device included in the information processing system 100 may be arranged in the same facility (for example, a research institute and a university) or in different facilities. Also, the number of each device included in the information processing system 100 is not particularly limited, and each device may be composed of a plurality of devices having the same function.

[0017] The imaging device 12 images the living tissue and the decellularized tissue and outputs the captured image. As the imaging device 12, for example, various microscopes such as a bright-field microscope, a scanning electron microscope (SEM: Scanning Electron Microscope), a transmission electron microscope (TEM: Transmission Electron Microscope), a fluorescence microscope, and a phase-contrast microscope can be applied.

[0018] For example, when using a bright-field microscope, stained images of living tissue and decellularized tissue are obtained. Figure 2 shows a schematic diagram of the stained image of living tissue before decellularization. Living tissue before decellularization contains ECM90 and cells containing cell nuclei92. By staining this living tissue with hematoxylin eosin (HE), for example, the cell nuclei92 are stained blue, and the tissue containing ECM90, cytoplasm, and intercellular matrix are stained in varying shades of pink. By photographing this stained living tissue with the imaging device 12, a living tissue image 50 similar to that shown in Figure 2 is obtained.

[0019] Figure 3 shows a schematic diagram of the stained image of decellularized tissue. In decellularized tissue, cells are removed from the living tissue, and ECM90 remains. By staining this decellularized tissue and photographing it with the imaging device 12, an image of decellularized tissue (second image 52) similar to that shown in Figure 3 is obtained.

[0020] There are various methods for decellularization, including chemical and physical methods. One example of a chemical method is the use of sodium dodecyl sulfate (SDS), a surfactant. One example of a physical method is the application of high hydrostatic pressure (HHP). It is also common practice to combine multiple methods as appropriate, for example, by combining SDS and HHP.

[0021] Currently, regardless of the decellularization method used, changes in the structure of the extracellular matrix (ECM) are unavoidable. Therefore, the decellularization process is optimized according to the properties of the source and recipient tissues (e.g., tissue density, cell density, fat content, and structure), as well as the characteristics and applications required for the final product.

[0022] For this optimization, it is important to evaluate the structural changes of the ECM before and after decellularization. These structural changes can be evaluated by observing images of each tissue before and after decellularization. For example, the image of the decellularized tissue in Figure 3 shows larger gaps between ECM90 cells compared to the image of the living tissue in Figure 2. Furthermore, the difference in staining intensity indicates a change in the density of some ECM90 cells. This change in ECM90 density may occur, for example, due to the treatment of the ECM90 on the surface of the living tissue during decellularization.

[0023] The information processing device 10 according to this embodiment enables objective evaluation by quantifying the structural changes of decellularized tissue based on images before and after decellularization. An example of the configuration of the information processing device 10 according to this embodiment will be described below.

[0024] Referring to Figure 4, an example of the hardware configuration of the information processing device 10 according to this embodiment will be described. The information processing device 10 includes a CPU (Central Processing Unit) 21, a non-volatile storage unit 22, and a memory 23 as a temporary storage area. The information processing device 10 also includes a display 24, an input unit 25, and a communication interface 26. The CPU 21, storage unit 22, memory 23, display 24, input unit 25, and communication interface 26 are connected to each other via a bus 28, such as a system bus and a control bus, enabling the exchange of various types of information.

[0025] The storage unit 22 is implemented by a storage medium such as an HDD (Hard Disk Drive), SSD (Solid State Drive), and flash memory. The information processing program 27 of the information processing device 10 is stored in the storage unit 22. The CPU 21 reads the information processing program 27 from the storage unit 22, expands it into memory 23, and executes the expanded information processing program 27. The CPU 21 is an example of the processor of this disclosure.

[0026] The display 24 is, for example, a liquid crystal display and displays various information. The input unit 25 includes a pointing device such as a mouse and a keyboard, and is used to input various information to the device. The display 24 may be configured as a touch panel and used in conjunction with the input unit 25.

[0027] Communication I / F26 is an interface for communicating with external devices. For this communication, wired communication standards such as Ethernet® or FDDI (Fiber Distributed Data Interface), or wireless communication standards such as 4G, 5G, or Wi-Fi® can be used. The information processing device 10 can appropriately include, for example, a server computer, a personal computer, a smartphone, a tablet terminal, and a wearable terminal.

[0028] Referring to Figure 1, an example of the functional configuration of the information processing device 10 according to this embodiment will be described. The information processing device 10 includes, as a functional configuration, an acquisition unit 30, a generation unit 31, a construction unit 32, a calculation unit 33, an output unit 34, and a correction unit 35. The CPU 21 executes the information processing program 27, thereby enabling the acquisition unit 30, generation unit 31, construction unit 32, calculation unit 33, output unit 34, and correction unit 35 to function.

[0029] (Construction of a structure map) First, the information processing device 10 constructs a structural map for quantitatively evaluating the structural changes of the decellularized tissue. The method for constructing the structural map is described below.

[0030] The acquisition unit 30 acquires a first image 51, which is an image of biological tissue captured by the imaging device 12. This first image 51 is the reference image for the structural map. As the first image 51, either an image of the biological tissue before decellularization or an image of the decellularized tissue may be used, but it is particularly preferable to use an image of the biological tissue before decellularization.

[0031] As mentioned above, decellularization causes various structural changes in the extracellular matrix (ECM) depending on the method used. While the structural map maps images of the decellularized tissue being evaluated, the nature of these structural changes is unknown. Therefore, it is desirable to use an image that mimics an ideal decellularized tissue with no changes in the ECM structure as the first image 51 that serves as the basis for the structural map.

[0032] For such a first image 51, a stained image of living tissue before decellularization is preferred. In particular, eosin staining (HE) is used, which stains only the tissue, cytoplasm, and intercellular matrix containing ECM90, without staining the cell nucleus. nuclei- By applying staining to living tissue, a stained image in which the cell nucleus is not visible can be obtained. This stained image without the cell nucleus is suitable for data expansion into the pseudo-image 56 described later.

[0033] The generation unit 31 generates a pseudo-image 56 by applying predetermined image processing to the first image 51. The pseudo-image 56 is an image added to accurately reduce the dimensionality when constructing a structure map using a so-called data augmentation technique. Specifically, it is preferable that the generation unit 31 generates multiple pseudo-images 56 by applying multiple different image processing to the first image 51.

[0034] Figure 5 shows an example of the first image 51 and three pseudo-images 56A to 56C generated based on it. For example, the generation unit 31 may perform at least one of the following image processing processes: binarization, sharpening, and tone curve adjustment. As an example, binarization can leave only the thicker fibers. By combining sharpening and binarization, the fibers can be made clearer. The color can be lightened by an inverted S-shaped tone curve adjustment.

[0035] These image processing techniques can simulate structural changes in images caused by decellularization, such as an increase in voids and changes in ECM90 staining density. By constructing a structural map using this simulated image 56, it is possible to construct a structural map that is not dependent on structural changes specific to a particular decellularization method, compared to using images of actual decellularized tissue obtained by existing decellularization methods. In other words, the images to be evaluated can be evaluated within a standardized set of standards that includes unified structural changes obtained through image processing, rather than within a standard set of standards that includes structural changes caused by existing decellularization methods. Therefore, it becomes possible to evaluate various types of decellularized tissues that exhibit different structural changes more objectively.

[0036] Furthermore, if the first image 51 is a stained image of biological tissue in which the cell nuclei are not visible due to eosin staining, a pseudo-image 56 without cell nuclei can be easily generated by the various image processing methods described above. On the other hand, if the first image 51 is a stained image of biological tissue in which the cell nuclei are visible due to HE staining, the generation unit 31 may perform image processing to remove the cell nuclei. Examples of such image processing include removing only the blue color of the cell nuclei, and removing the cell nuclei by pattern matching based on the shape of the cell nuclei.

[0037] The construction unit 32 constructs a structure map showing the distribution of features from the first image 51 and the pseudo-image 56. Figure 6 shows an example of the structure map 40. In Figure 6, group U is the set of features extracted from the first image 51. Groups UA to UC are the set of features extracted from the pseudo-images 56A to 56C, respectively. In this way, when there are multiple pseudo-images 56, the construction unit 32 constructs a structure map showing the distribution of features from the first image 51 and the multiple pseudo-images 56.

[0038] Furthermore, it is preferable for the construction unit 32 to divide each of the first image 51 and the pseudo-image 56 into multiple regions and construct a structure map showing the distribution of feature quantities for each region. The construction of the structure map 40 itself is possible if at least one feature quantity (sample point) can be plotted from each image, but the accuracy of the analysis can be improved as the number of plots increases. This is because increasing the number of sample points can reduce bias caused by differences in tissue structure between different parts, as well as individual differences between living tissue and decellularized tissue. In Figure 6, each of the sample points indicated by white circles corresponds to a feature quantity of one region in the first image 51 and the pseudo-image 56. For example, group U is a set of feature quantities for each region of the first image 51.

[0039] Specifically, the construction unit 32 extracts features from the first image 51 and the pseudo-image 56 using an extractor that extracts features from the input image (or region). As the extractor, known technologies such as a convolutional neural network (CNN) and the encoder part of an autoencoder can be applied. As for CNNs, known technologies such as AlexNet, InceptionV3, Resnet50, and Xception can be applied.

[0040] Features are represented by a set of one or more numerical values, i.e., n-dimensional (n is 1 or greater) vector data. As an example, Figure 6 shows a structure map 40 of two-dimensional features Y1 and Y2. Note that the output of the extractor may have large dimensionality numbers, such as 512, 1024, and 2048. Therefore, the construction unit 32 may perform dimensionality reduction on the features. Known techniques such as Principal Component Analysis (PCA), t-SNE (t-distributed Stochastic Neighbor Embedding), and UMAP (Uniform Manifold Approximation and Projection) can be applied as dimensionality reduction methods.

[0041] (Quantitative evaluation of structural changes) The information processing device 10 quantitatively evaluates the structural changes of any decellularized tissue using the structural map constructed as described above. The quantitative evaluation method for structural changes is described below.

[0042] The acquisition unit 30 acquires a second image 52, which is an image of decellularized tissue. The second image 52 may be taken in the same environment as the first image 51, or in a different environment, but it is the same type of image as the first image 51. For example, if the first image 51 is a stained image of living tissue, then the second image 52 is also a stained image of decellularized tissue.

[0043] Furthermore, the acquisition unit 30 acquires a third image, which is an image of the living tissue before the cellular components are removed from the decellularized tissue that is the subject of the second image 52. Preferably, the third image is taken in the same environment as the second image 52 and is of the same type as the first image 51 and the second image 52.

[0044] The calculation unit 33 maps the features of the second image 52 and the features of the third image to the structure map 40, respectively. Figure 7 shows an example in which the features of the two second images 52-1 and 52-2 and the third image are mapped to the structure map 40 in Figure 6. In Figure 7, group U0 is the set of features extracted from the third image. Group U1 is the set of features extracted from the first second image 52-1. Group U2 is the set of features extracted from the second second image 52-2. Hereafter, when the two second images 52-1 and 52-2 are not distinguished, they will simply be referred to as second image 52.

[0045] Specifically, the calculation unit 33, similar to the construction unit 32, divides the second image 52 into multiple regions and maps the features of each region to the structure map 40. In Figure 7, each of the sample points indicated by rectangles corresponds to the features of one region in the second image 52-1. Each of the sample points indicated by triangles corresponds to the features of one region in the second image 52-2. Feature extraction and dimensionality reduction are performed in the same manner as in the construction unit 32.

[0046] Similarly, the calculation unit 33 divides the third image into multiple regions and maps the features of each region to the structure map 40. In Figure 7, each of the sample points indicated by black circles corresponds to a feature of one region in the third image. Feature extraction and dimensionality reduction are performed in the same manner as in the construction unit 32.

[0047] Next, the calculation unit 33 determines a reference point based on the features of the third image in the structure map. For example, the calculation unit 33 may calculate the average value for each variable from the distribution of features of the third image (group U0) and determine that average value as the reference point. This reference point determined based on the features of the third image is an example of a "predetermined reference point" in this disclosure.

[0048] Furthermore, the calculation unit 33 calculates the feature quantities of the second image 52 in the structure map 40 and the distance scale from the reference point. Specifically, the calculation unit 33 calculates at least one of the feature quantities for each region of the second image 52 in the structure map 40 and the distance scale from the reference point.

[0049] As an example of a distance measure in this case, we will explain how to calculate the Mahalanobis distance. The Mahalanobis distance is a distance measure between a sample point and a distribution. The Mahalanobis distance d from a vector y representing a sample point to a distribution with a mean vector μ and a covariance matrix Σ is expressed by the following equation (1). The vector y is represented by a vector (y1, y2, ..., yn) consisting of n variables. The mean vector μ is represented by a vector (μ1, μ2, ..., μn) representing the mean values ​​of each variable in the distribution. The covariance matrix Σ is a matrix that arranges the variances and covariances between each variable in the distribution. (y-μ) T This is the transpose matrix of (y-μ).

[0050]

number

[0051] The calculation unit 33 calculates the mean value for each variable from the feature distribution of the third image (group U0), and sets the vector consisting of these mean values ​​as vector y. As described above, this is the vector representing the reference point. The calculation unit 33 also calculates the mean vector μ and the covariance matrix Σ from the feature distribution of the second image 52-1 (group U1). By calculating equation (1) using these, the Mahalanobis distance d1 from the reference point to the feature distribution of the second image 52-1 is obtained.

[0052] Similarly, the calculation unit 33 calculates the mean vector μ and the covariance matrix Σ from the feature distribution (group U2) of the second image 52-2. By calculating equation (1) using these and vector y, the Mahalanobis distance d2 from the reference point to the feature distribution of the second image 52-2 is obtained.

[0053] The distance scale using the Mahalanobis distance described above is just one example, and various modifications are possible. For example, the calculation unit 33 may use the centroid instead of the mean in the calculation process of the Mahalanobis distance. The calculation unit 33 may also normalize the distance scale or perform other operations to make it easier for the user to interpret and compare changes in the structure of the ECM.

[0054] For example, the calculation unit 33 may apply any distance measure other than the Mahalanobis distance, such as the Euclidean distance, Manhattan distance, and Chebyshev distance (maximum distance). When using these distance measures, for example, the mean or centroid distance of the distributions, the shortest distance between distributions, and the longest distance between distributions may be applied. For example, the calculation unit 33 may calculate the vector y of the reference point and the Euclidean distance to each feature (each sample point) of the second image 52, and use representative values ​​such as the mean, minimum, and maximum values ​​of these Euclidean distances as distance measures for the second image 52.

[0055] The output unit 34 outputs the calculated distance scale as a numerical value indicating the degree of structural change in the decellularized tissue. Figure 8 shows an example of screen D output to the display 24 by the output unit 34. Screen D displays the second images 52-1 and 52-2, which are images of the two decellularized tissues being evaluated. In addition, the Mahalanobis distance, as an example of the distance scale calculated for each divided region, is displayed as a heat map of "deviation" overlaid on each image. Such a heat map makes it possible to visualize the variability of structural changes in the tissue in each image. Furthermore, for each image, the average value of the Mahalanobis distance calculated for each divided region is displayed as the "average deviation".

[0056] (Image pre-correction) Incidentally, the first image 51, the second image 52, and the third image may look different depending on the shooting environment, staining environment, and the skill level of the experimenter, even if they are actually in the same state. Therefore, it is preferable for the information processing device 10 to perform correction processing on each of the first image 51, the second image 52, and the third image before extracting features and mapping them to the structure map 40. The following describes the method for this pre-correction of images.

[0057] First, the correction unit 35 determines the parameters for the correction process based on a histogram of pixel values ​​for each pixel of the first image 51. Pixel values ​​include, for example, saturation, brightness, and hue. Specifically, it is preferable that the correction unit 35 uses at least one of saturation and brightness as the pixel values ​​referenced to determine the parameters for the correction process.

[0058] Figure 9 shows the first images 51 and 51P before and after correction, along with their saturation and brightness histograms. For example, the correction unit 35 obtains the corrected first image 51P by adjusting the white balance so that the empty space in the first image 51 where ECM 90 does not exist becomes the whitest (brightest). As a result, the range of the saturation histogram, which was 0 for the minimum value and 255 for the maximum value before correction, is expanded to S1 for the minimum value and S2 for the maximum value after correction (S1 and S2 are between 0 and 255). Similarly, the range of the brightness histogram, which was 0 for the minimum value and 255 for the maximum value before correction, is expanded to B1 for the minimum value and B2 for the maximum value after correction (B1 and B2 are between 0 and 255).

[0059] The correction unit 35 determines the minimum and maximum values ​​of the corrected saturation, as well as the minimum and maximum values ​​of the corrected brightness, as parameters for the correction process. Then, the correction unit 35 applies the correction process to the second image 52 and the third image using the determined parameters.

[0060] Figure 10 shows the second images 52 and 52P before and after correction, along with their saturation and brightness histograms. As shown in Figure 10, the range of the saturation histogram for the corrected second image 52P is the same as for the corrected first image 51P, with a minimum value of S1 and a maximum value of S2. Similarly, the range of the brightness histogram for the corrected second image 52P is the same as for the corrected first image 51P, with a minimum value of B1 and a maximum value of B2.

[0061] Specifically, the correction unit 35 applies correction processing to the first image 51, the second image 52, and the third image using the same parameters. The generation unit 31 generates a pseudo-image 56 using the first image 51P after correction processing. The construction unit 32 constructs a structure map 40 using the first image 51P after correction processing. The calculation unit 33 performs mapping to the structure map and calculates the distance scale using the second image 52P and the third image after correction processing. Through the above correction processing, the influence on the distance scale caused by apparent differences in the images can be suppressed.

[0062] Next, the operation of the information processing device 10 will be explained with reference to Figure 11. In the information processing device 10, the CPU 21 executes the information processing program 27, thereby executing the information processing shown in Figure 11. This information processing is executed, for example, when the user issues an instruction to start execution.

[0063] In step S10, the acquisition unit 30 acquires a first image, which is an image of biological tissue. In step S12, the generation unit 31 generates a pseudo-image by applying predetermined image processing to the first image acquired in step S10. In step S14, the construction unit 32 extracts feature quantities from the first image acquired in step S10 and the pseudo-image generated in step S12, and constructs a structure map showing the distribution of feature quantities.

[0064] In step S16, the acquisition unit 30 acquires a second image, which is an image of decellularized tissue. In step S18, the calculation unit 33 extracts feature quantities from the second image acquired in step S16 and maps the extracted feature quantities to the structure map constructed in step S14. In step S20, the calculation unit 33 calculates the distance scale between the feature quantities of the second image in the structure map and predetermined reference points in the structure map. In step S22, the output unit 34 outputs the distance scale calculated in step S20 and terminates this information processing.

[0065] As described above, the information processing device 10 according to this embodiment includes a processor. The processor acquires a first image, which is an image of biological tissue, generates a pseudo-image by applying predetermined image processing to the first image, and constructs a structure map showing the distribution of feature quantities of the first image and the pseudo-image, respectively. The processor also acquires a second image, which is an image of decellularized tissue from which cellular components have been removed from biological tissue, maps the feature quantities of the second image to the structure map, and calculates a distance scale between the feature quantities of the second image in the structure map and predetermined reference points in the structure map.

[0066] In other words, the information processing device 10 constructs a structural map that does not include the effects of decellularization on the biological tissue or differences in the imaging environment, by using an image of the biological tissue that has not undergone decellularization (first image) and a pseudo-image generated based on it. Furthermore, the information processing device 10 quantifies the extent to which the tissue structure has changed due to decellularization by calculating a distance scale between the decellularized image (second image) and a reference point based on this structural map.

[0067] Therefore, according to the information processing device 10 of this embodiment, it becomes possible to objectively evaluate the structural changes of decellularized tissue for each decellularization method. This facilitates the interpretation and comparison of each decellularization method, contributing to optimization. Ultimately, this can promote the research and development of decellularized tissue for transplantation and regenerative medicine.

[0068] In the above embodiment, a configuration in which stained images are used as the first image 51 and the second image 52 has been described, but the invention is not limited to this. For example, unstained images of living tissue and decellularized tissue obtained by using a phase-contrast microscope may be used as the first image 51 and the second image 52. Alternatively, for example, images obtained by Fourier transforming a stained image (or unstained image) may be used as the first image 51 and the second image 52.

[0069] Furthermore, although the above embodiment describes a configuration in which the reference points used for calculating the distance scale are determined based on the features of the third image, the invention is not limited to this configuration. For example, the calculation unit 33 may calculate the reference points based on the features of the first image 51 instead of the features of the third image. Alternatively, the calculation unit 33 may calculate the origin of the structure map, as well as representative values ​​such as the mean, minimum, and maximum values ​​of all sample points included in the structure map, as reference points. In these cases, the third image is not required for calculating the distance scale, so the input of the third image may be omitted.

[0070] (Example 1) Figure 12 shows an example of a structural map constructed using the technology of this disclosure. In this example, the first image, which serves as the base for the structural map, is an eosin-stained image of the aorta before decellularization. The second image, which is the subject of evaluation, is an eosin-stained image of the decellularized aorta obtained by the HHP and SDS methods, respectively. The third image is an eosin-stained image of the aorta before decellularization was performed on the decellularized tissue that was the subject of the second image. The shooting environment, staining environment, and experimenter were the same for the first, second, and third images.

[0071] In this embodiment, a structure map was constructed based on the first image and seven pseudo-images generated from the first image. The pseudo-images used were those obtained by applying binarization, sharpening, and tone curve adjustment to the first image, respectively, as well as those obtained by stretching and deforming the entire set of pseudo-images by 1.5 times, and those obtained by stretching and deforming the entire first image by 1.5 times.

[0072] Each image, excluding the stretched and deformed image, was 1480 x 1960 pixels in size. These were divided into 100 x 100 pixel regions, and features were plotted for each region. InceptionV3 was used as the feature extractor. PCA was used as the dimensionality reduction method.

[0073] Figure 12 shows that the direction of feature distribution differs between HHP and SDS. In other words, the way in which structural changes occur differs depending on the decellularization method, and this can be evaluated from the images. Furthermore, the degree of overlap in the feature distribution of "Image 1" and "Image 3," both obtained by imaging the aorta before decellularization, shows that if there is no structural change, the feature distribution will be similar. From the above, it is thought that mapping decellularized tissue prepared by the novel method onto existing structural maps will function effectively.

[0074] Furthermore, it was found that the directionality of the feature distribution differed significantly in each pseudo-image that had undergone binarization, sharpening, and tone curve adjustment. In other words, these image processing methods were able to comprehensively represent the ways in which the structure of various tissues changes, demonstrating the effectiveness of constructing structure maps using pseudo-images.

[0075] On the other hand, in this embodiment, it was found that the distribution of features in each stretched pseudo-image was similar to the distribution of features in each pseudo-image from which it was deformed. In other words, the pseudo-representation of gap expansion due to stretching deformation did not easily show differences in features compared to other image processing such as binarization. Therefore, in this embodiment, at least one of binarization, sharpening, and tone curve adjustment is considered suitable as the image processing for generating pseudo-images. However, this embodiment is not limited to these image processing methods, and it is preferable to appropriately apply various image processing methods, including stretching deformation, that are effective in constructing a structure map, depending on the dimensionality reduction method and feature extractor, for example.

[0076] (Example 2) Figure 13 shows another example of a structure map constructed using the technology of this disclosure. In this example, we confirmed how structural changes in response to decellularization processing time can be evaluated using a structure map. The first image, which forms the basis of the structure map, and the third image, which is used to calculate the distance scale, are stained images of the tissue before decellularization using eosin staining. As the second image to be evaluated, stained images using eosin staining were obtained while varying the decellularization processing time using the SDS method to 1 day, 3 days, and 5 days. The shooting environment, staining environment, and experimenter were the same for the first, second, and third images.

[0077] In this example, a structure map was constructed based on the first image and three pseudo-images generated from the first image. The pseudo-images were created by applying binarization, sharpening, and tone curve adjustment to the first image, respectively. Each image was 1480 × 1960 pixels in size, and this was divided into 100 × 100 pixel regions, with features plotted for each region. Xception was used as the feature extractor. PCA was used as the dimensionality reduction method.

[0078] Figure 13 shows that the features of the second image, which had a longer decellularization processing time, were plotted further away from the features of the third image. In other words, it was shown that the longer the decellularization processing time, the more the structure of the ECM changes, and this can be evaluated using the distance scale (degree of deviation) calculated with the structure map. It is generally known that the structure of the ECM tends to change with longer decellularization processing times, and the results of this example were found to be consistent with this trend.

[0079] In this embodiment, each process is executed on any computer. Furthermore, any computer may execute these processes using a processor as hardware, a program as software, or a combination thereof. In that case, the processor is configured to work in cooperation with the program to execute the various processes in this embodiment, and can function as a unit or means in this embodiment. Also, the execution order of the processes by the processor is not limited to the order described and may be changed as appropriate. Any computer may be a general-purpose computer, a computer designed for a specific purpose, a workstation, or any other system capable of executing each process.

[0080] A processor may consist of one or more hardware components, and the type of hardware is not limited. For example, a processor may consist of a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a programmable logic device such as an FPGA (Field Programmable Gate Array), a dedicated circuit for executing a specific process such as an ASIC (Application Specific Integrated Circuit), a GPU (Graphic Processing Unit), or an NPU (Neural Processing Unit). Furthermore, the type of hardware may be a combination of different types of hardware. When multiple hardware components are configured to execute one or more processes of a processor, these components may reside in physically separate devices or in the same device. Also, in any embodiment, the order of each process performed by the processor is not limited to the order described above and may be changed as appropriate. Hardware is composed of electrical circuits (circuitry) that combine circuit elements such as semiconductor elements.

[0081] Furthermore, the program may be firmware or software such as microcode. Alternatively, the program may be, for example, a set of program modules, each function of which may be implemented by a processor configured to perform its respective function. The program may be program code or multiple code segments stored on one or more non-temporary computer-readable media (e.g., storage media or other storage). The program may be divided and stored on multiple non-temporary computer-readable media located in physically separate devices. Program code or code segments may represent any combination of procedures, functions, subprograms, routines, subroutines, modules, software packages, classes, or instructions, data structures, or program statements. Program code or code segments may be connected to other code segments or hardware circuits by sending and receiving information, data, arguments, parameters, or memory contents.

[0082] Furthermore, although the above embodiment describes an embodiment in which the information processing program 27 is pre-stored (installed) in the storage unit 22, the invention is not limited to this. The information processing program 27 may be provided in the form of a recording medium such as a CD-ROM (Compact Disc Read Only Memory), DVD-ROM (Digital Versatile Disc Read Only Memory), or USB (Universal Serial Bus) memory. Alternatively, the information processing program 27 may be provided in the form of a download from an external device via a network.

[0083] The technology disclosed herein extends to all program products. A program product includes all forms of products for providing programs. For example, a program product includes programs provided via a network such as the Internet, as well as non-temporary computer-readable recording media such as CD-ROMs, DVD-ROMs, and USB memory sticks on which programs are stored.

[0084] The technology of this disclosure can also be appropriately combined with the above-described embodiments and modifications. The descriptions and illustrations shown above are detailed explanations of the parts relating to the technology of this disclosure and are merely examples of the technology of this disclosure. For example, the above descriptions of the configuration, function, operation, and effect are examples of the configuration, function, operation, and effect of the parts relating to the technology of this disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements added, or replaced from the descriptions and illustrations shown above, as long as they do not deviate from the spirit of the technology of this disclosure.

[0085] The following additional information is disclosed regarding the above embodiments. [Note 1] The processor comprises, We obtain the first image, which is an image of living tissue. A pseudo-image is generated by applying predetermined image processing to the first image. A structure map is constructed showing the distribution of feature quantities for the first image and the pseudo-image, A second image is obtained, which is an image of decellularized tissue from which cellular components have been removed from living tissue. The features of the second image are mapped to the structure map, The distance scale between the feature quantities of the second image in the structure map and a predetermined reference point in the structure map is calculated. Information processing device. [Note 2] The aforementioned processor, Multiple pseudo-images are generated by applying multiple different image processing steps to the first image. Construct the structure map that shows the distribution of feature quantities for the first image and the plurality of pseudo-images. The information processing device described in Appendix 1. [Note 3] The processor performs at least one of the following image processing steps: binarization, sharpening, and tone curve adjustment. The information processing device described in Appendix 1 or Appendix 2. [Note 4] The aforementioned processor, The first image and the pseudo-image are each divided into multiple regions, and a structure map is constructed that shows the distribution of feature quantities for each region. The second image is divided into multiple regions, and the feature quantities for each region are mapped to the structure map. The distance scale is calculated between at least one of the feature quantities for each region of the second image in the structure map and the reference point. An information processing device as described in any one of the appendices 1 to 3. [Note 5] The aforementioned processor, A third image is obtained, which is an image of the living tissue before the cellular components are removed from the decellularized tissue. The features of the third image are mapped to the structure map, Based on the feature quantities of the third image in the structure map, the reference point is determined. An information processing device as described in any one of the appendices 1 through 4. [Note 6] The first image above is a stained image of the biological tissue, The second image above is a stained image of the decellularized tissue, The aforementioned processor, Based on the histogram of pixel values ​​for each pixel in the first image, the parameters for the correction process are determined. The correction process using the same parameters is applied to each of the first and second images. Using the first image after the correction process, the pseudo-image is generated and the structure map is constructed. Using the second image after the correction process, mapping to the structure map and calculation of the distance scale are performed. An information processing device as described in any one of the appendices 1 through 5. [Note 7] The processor uses at least one of saturation and brightness as the pixel value. The information processing device described in Appendix 6. [Note 8] The processor calculates the Mahalanobis distance as the distance measure. An information processing device as described in any one of the appendices 1 through 7. [Note 9] We obtain the first image, which is an image of living tissue. A pseudo-image is generated by applying predetermined image processing to the first image. A structure map is constructed showing the distribution of feature quantities for the first image and the pseudo-image, A second image is obtained, which is an image of decellularized tissue from which cellular components have been removed from living tissue. The features of the second image are mapped to the structure map, The distance scale between the feature quantities of the second image in the structure map and a predetermined reference point in the structure map is calculated. An information processing method in which a computer performs the processing. [Note 10] We obtain the first image, which is an image of living tissue. A pseudo-image is generated by applying predetermined image processing to the first image. A structure map is constructed showing the distribution of feature quantities for the first image and the pseudo-image, A second image is obtained, which is an image of decellularized tissue from which cellular components have been removed from living tissue. The features of the second image are mapped to the structure map, The distance scale between the feature quantities of the second image in the structure map and a predetermined reference point in the structure map is calculated. An information processing program that instructs a computer to perform a task. [Explanation of symbols]

[0086] 10 Information Processing Devices 12. Imaging device 21 CPU 22 Memory section 23 memory 24 displays 25 Input section 26 Communication I / F 27 Information Processing Programs 28 buses 30 Acquisition Department 31 Generation part 32 Construction Department 33 Calculation Section 34 Output section 35 Correction section 40 Structural Map 50 images of biological tissue 51, 51P First image 52, 52P, 52-1, 52-2 Second image 56, 56A~56C pseudo image 90 ECM 92 cell nucleus 100 Information Processing Systems d1, d2 Mahalanobis distance D screen U, U0~U2, UA~UC Group

Claims

1. The processor comprises, We obtain the first image, which is an image of living tissue. A pseudo-image is generated by applying predetermined image processing to the first image. A structure map is constructed showing the distribution of feature quantities for the first image and the pseudo-image, A second image is obtained, which is an image of decellularized tissue from which cellular components have been removed from living tissue. The feature quantities of the second image are mapped to the structure map, The distance scale between the feature quantities of the second image in the structure map and a predetermined reference point in the structure map is calculated. Information processing device.

2. The aforementioned processor, By applying a plurality of different image processing operations to the first image, a plurality of pseudo-images are generated. Construct the structure map that shows the distribution of feature quantities for the first image and the plurality of pseudo-images. The information processing apparatus according to claim 1.

3. The processor performs at least one of the following image processing steps: binarization, sharpening, and tone curve adjustment. The information processing apparatus according to claim 1.

4. The aforementioned processor, The first image and the pseudo-image are each divided into multiple regions, and a structure map is constructed that shows the distribution of feature quantities for each region. The second image is divided into multiple regions, and the feature quantities for each region are mapped to the structure map. The distance scale between at least one of the feature quantities for each region of the second image in the structure map and the reference point is calculated. The information processing apparatus according to claim 1.

5. The aforementioned processor, A third image is obtained, which is an image of the living tissue before the cellular components are removed from the decellularized tissue. The feature quantities of the third image are mapped to the structure map, Based on the feature quantities of the third image in the structure map, the reference point is determined. The information processing apparatus according to claim 1.

6. The first image is a stained image of the biological tissue, The second image above is a stained image of the decellularized tissue, The aforementioned processor, Based on the histogram of pixel values ​​for each pixel in the first image, the parameters for the correction process are determined. The correction process using the same parameters is applied to each of the first and second images. Using the first image after the correction process, the pseudo-image is generated and the structure map is constructed. Using the second image after the correction process, mapping to the structure map and calculation of the distance scale are performed. The information processing apparatus according to claim 1.

7. The processor uses at least one of saturation and brightness as the pixel value. The information processing apparatus according to claim 6.

8. The processor calculates the Mahalanobis distance as the distance measure. The information processing apparatus according to claim 1.

9. We obtain the first image, which is an image of living tissue. A pseudo-image is generated by applying predetermined image processing to the first image. A structure map is constructed showing the distribution of feature quantities for the first image and the pseudo-image, A second image is obtained, which is an image of decellularized tissue from which cellular components have been removed from living tissue. The feature quantities of the second image are mapped to the structure map, The distance scale between the feature quantities of the second image in the structure map and a predetermined reference point in the structure map is calculated. An information processing method in which a computer performs the processing.

10. We obtain the first image, which is an image of living tissue. A pseudo-image is generated by applying predetermined image processing to the first image. A structure map is constructed showing the distribution of feature quantities for the first image and the pseudo-image, A second image is obtained, which is an image of decellularized tissue from which cellular components have been removed from living tissue. The feature quantities of the second image are mapped to the structure map, The distance scale between the feature quantities of the second image in the structure map and a predetermined reference point in the structure map is calculated. An information processing program that instructs a computer to perform a task.

Images