Pathological diagnosis support device, pathological diagnosis support method, and program

Abstract

The present invention provides a pathological diagnosis support device that comprises: a determination unit that determines whether or not correction is necessary to optimize a classification result in intraoperative rapid cytodiagnosis, said correction being correction of a parameter indicating an effective magnification, which is the magnification of the size of a subject in an image of interest relative to the actual size of the subject, the image of interest having been obtained by using a camera to capture a microscope-produced image of a specimen cell to be classified as benign or malignant in intraoperative rapid cytodiagnosis; a notification unit that issues a notification indicating that parameter correction is necessary when correction is determined to be necessary; and a classification unit that classifies the specimen cell as benign or malignant when correction is determined to be unnecessary.

Description

Pathological diagnosis support device, pathological diagnosis support method, and program

[0001] The present disclosure relates to a pathological diagnosis support device, a pathological diagnosis support method, and a program.

[0002] 2. Description of the Related Art There are known techniques for classifying cells included as subjects in an image as benign or malignant.

[0003] For example, Patent Literature 1 discloses a classification device that classifies cells contained as subjects in an image that includes a portion of a pathology sample as a subject as to whether they are benign or malignant. The classification device adjusts the imaging magnification of the image depending on the number of cells contained as subjects in the image.

[0004] International Publication No. 2023-105548

[0005] In AI diagnostic support for rapid on-site evaluation (ROSE), an image containing a group of specimen cells, including at least one specimen cell to be classified as benign or malignant, is captured using a camera installed in a microscope.

[0006] In this configuration, even if images of the same subject are taken, there is a possibility that the size of the subject in the image may differ. For example, when the same subject is taken using two different microscopes, even if the same pathological sample is taken using the same camera and the same nominal objective lens in each of the two different microscopes, the distance from the camera focus to the subject may be different, and therefore the size of the subject in the image may differ.

[0007] Furthermore, even in multiple images taken using the same microscope, camera, and objective lens, the distance from the camera's focal point to the subject may change over time or due to physical impact while the microscope is in use, which may result in differences in the size of the subject in the images, even when the same subject is photographed using the same microscope, camera, and objective lens.

[0008] Thus, even if the same subject is captured in multiple images, the size of the image may differ. That is, even if a subject is captured using an objective lens with the same nominal value in a microscope equipped with a camera, the magnification (effective magnification) of the captured image may differ. If a sample cell group is classified as benign or malignant using multiple images captured at different effective magnifications, the accuracy of the classification may be reduced.

[0009] The present disclosure has been made in consideration of the above-mentioned problems, and one exemplary purpose thereof is to provide a technology for improving the accuracy of classifying a sample cell group into benign and malignant in pathological diagnosis.

[0010] A pathology diagnosis support device according to an exemplary aspect of the present disclosure includes an acquisition means for acquiring a target image obtained by capturing an image of specimen cells to be classified as benign or malignant under a microscope using a camera, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination means for determining whether correction of the parameter is necessary; a notification means for notifying the user that correction of the parameter is necessary when the determination means determines that correction is necessary; and a classification means for classifying the specimen cells as benign or malignant using the parameter when the determination means determines that correction is not necessary.

[0011] A pathology diagnosis support method according to an exemplary aspect of the present disclosure includes an acquisition process in which at least one processor acquires a target image obtained by a camera capturing an image of specimen cells to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination process in which the parameter is determined to be corrected; a notification process in which, if the determination process determines that correction is necessary, the processor notifies the user that the parameter needs to be corrected; and, if the determination process determines that correction is not necessary, a classification process in which the parameter is used to classify the specimen cells as benign or malignant.

[0012] A program according to an exemplary aspect of the present disclosure is a program that causes a computer to function as a pathology diagnosis support device, and the program causes the computer to execute the following processes: an acquisition process that acquires a target image obtained by capturing an image of specimen cells to be classified as benign or malignant under a microscope using a camera, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination process that determines whether correction of the parameter is necessary; a notification process that notifies the user that correction of the parameter is necessary if it is determined in the determination process that correction is necessary; and a classification process that uses the parameter to classify the specimen cells as benign or malignant if it is determined in the determination process that correction is not necessary.

[0013] According to one exemplary aspect of the present disclosure, an exemplary effect is achieved in that a technology can be provided that improves the accuracy of classifying specimen cells into benign and malignant in pathological diagnosis.

[0014] FIG. 1 is a block diagram showing the configuration of a pathology diagnosis support device according to the present disclosure. FIG. 2 is a flow diagram showing the flow of a pathology diagnosis support method according to the present disclosure. FIG. 3 is a diagram showing an example of pathology diagnosis using a pathology diagnosis support device according to the present disclosure. FIG. 4 is a block diagram showing the configuration of a pathology diagnosis support device according to the present disclosure. FIG. 5 is a flow diagram showing the flow of a pathology diagnosis support method according to the present disclosure. FIG. 6 is a flow diagram showing the flow of processing performed by a determination unit according to the present disclosure. FIG. 7 is a diagram showing an example of processing by a calculation unit according to the present disclosure to calculate similarity. FIG. 8 is a flow diagram showing the flow of processing performed by a correction unit according to the present disclosure. FIG. 9 is an example of an image processed by a correction unit according to the present disclosure. FIG. 10 is a block diagram showing the configuration of a computer that functions as a pathology diagnosis support device according to the present disclosure.

[0015] The following are examples of embodiments of the present invention. However, the present invention is not limited to the exemplary embodiments shown below, and various modifications are possible within the scope of the claims. For example, embodiments obtained by appropriately combining the technical means employed in the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, embodiments obtained by appropriately omitting some of the technical means employed in the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, the effects mentioned in the exemplary embodiments shown below are examples of effects expected in the exemplary embodiments, and do not define the scope of the present invention. In other words, embodiments that do not exhibit the effects mentioned in the exemplary embodiments shown below may also be included in the scope of the present invention.

[0016] [First Exemplary Embodiment] A first exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. This exemplary embodiment is a basic form for each of the exemplary embodiments described below. Note that the scope of application of each technical means employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technical means employed in this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical obstacles arise. Furthermore, each technical means shown in the drawings referenced to explain this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical obstacles arise.

[0017] (Configuration of Pathological Diagnosis Support Device) The configuration of the pathological diagnosis support device 1 will be described with reference to Fig. 1. Fig. 1 is a block diagram showing the configuration of the pathological diagnosis support device 1. As shown in Fig. 1, the pathological diagnosis support device 1 includes an acquisition unit 11, a determination unit 12, a notification unit 13, and a classification unit 14. In this exemplary embodiment, the acquisition unit 11, the determination unit 12, the notification unit 13, and the classification unit 14 are components that respectively realize an acquisition means, a determination means, a notification means, and a classification means.

[0018] The acquisition unit 11 acquires a target image obtained by photographing a microscope image of specimen cells to be classified as benign or malignant in intraoperative rapid cytology, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject. The acquisition unit 11 supplies the acquired target image to the classification unit 14, and supplies the acquired parameter to the determination unit 12.

[0019] In addition, the acquisition unit 11 may be configured to acquire a target image captured by a camera of a group of specimen cells that are classified as benign or malignant under a microscope in intraoperative rapid cytology, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject.

[0020] The determination unit 12 determines whether or not correction of the parameters acquired by the acquisition unit 11 is necessary. The determination unit 12 supplies the determination result to the notification unit 13 and the classification unit 14.

[0021] When the determining unit 12 determines that correction is necessary, the notifying unit 13 notifies that parameter correction is necessary.

[0022] If the determining unit 12 determines that correction is not necessary, the classifying unit 14 uses the parameters to classify the specimen cells as benign or malignant.

[0023] In addition, when the acquisition unit 11 acquires a target image obtained by photographing a microscope image of a specimen cell group to be classified as benign or malignant in intraoperative rapid cytology, the classification unit 14 may use parameters to classify each specimen cell group as benign or malignant if the judgment unit 12 determines that correction is not necessary.

[0024] (Effects of the pathology diagnosis support device) As described above, the pathology diagnosis support device 1 employs a configuration including an acquisition unit 11 that acquires a target image obtained by a camera capturing an image of specimen cells to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination unit 12 that determines whether or not the parameter acquired by the acquisition unit 11 needs to be corrected; a notification unit 13 that notifies the fact that parameter correction is necessary when the determination unit 12 determines that correction is necessary; and a classification unit 14 that classifies the specimen cells as benign or malignant using the parameter when the determination unit 12 determines that correction is not necessary.

[0025] Therefore, the pathological diagnosis support device 1 has the effect of providing a technique for improving the accuracy in classifying specimen cells into benign and malignant.

[0026] In addition, the pathological diagnosis support device 1 may be configured to include an acquisition unit 11 that acquires a target image obtained by a camera of a microscope image of a specimen cell group to be classified as benign or malignant in intraoperative rapid cytology, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination unit 12 that determines whether or not correction of the parameter acquired by the acquisition unit 11 is necessary; a notification unit 13 that notifies the user that parameter correction is necessary if the determination unit 12 determines that correction is necessary; and a classification unit 14 that uses the parameter to classify each specimen cell group as benign or malignant if the determination unit 12 determines that correction is not necessary.

[0027] Therefore, the pathological diagnosis support device 1 has the effect of providing a technology that improves the accuracy of classifying individual specimen cell groups into benign and malignant.

[0028] (Flow of Pathological Diagnosis Support Method) The flow of pathological diagnosis support method S1 will be described with reference to Fig. 2. Fig. 2 is a flow chart showing the flow of pathological diagnosis support method S1. As shown in Fig. 2, pathological diagnosis support method S1 includes an acquisition process S11, a determination process S12, a notification process S14, and a classification process S15.

[0029] In the acquisition process S11, the acquisition unit 11 acquires a target image obtained by photographing a microscope image of specimen cells to be classified as benign or malignant in the intraoperative rapid cytology, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject. The acquisition unit 11 supplies the acquired target image to the classification unit 14 and supplies the acquired parameter to the determination unit 12.

[0030] In addition, in the acquisition process S11, the acquisition unit 11 may be configured to acquire a target image captured by a camera of a microscope image of a group of specimen cells classified as benign or malignant in intraoperative rapid cytology, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject.

[0031] (Determination Process S12) In the determination process S12, the determination unit 12 determines whether or not correction of the parameters acquired by the acquisition unit 11 is necessary. The determination unit 12 supplies the determination result to the notification unit 13 and the classification unit 14.

[0032] (Step S13) In step S13, the notification unit 13 and the classification unit 14 determine whether the determination unit 12 has determined that parameter correction is necessary.

[0033] (Notification Process S14) If the determination unit 12 determines in the notification process S13 that correction is necessary (step S13: YES), the notification unit 13 notifies the user in the notification process S14 that parameter correction is necessary.

[0034] (Classification Process S15) If the determination unit 12 determines in step S13 that correction is not required (step S13: NO), the classification unit 14 classifies the specimen cells into benign or malignant cells using the parameters in classification process S15.

[0035] In addition, in the acquisition process S11, when the acquisition unit 11 acquires a target image obtained by photographing with a camera an image of a specimen cell group that is to be classified as benign or malignant in intraoperative rapid cytology under a microscope, if the judgment unit 12 determines that correction is not necessary, in the classification process S15, the classification unit 14 may use parameters to classify each specimen cell group as benign or malignant.

[0036] (Effects of Pathological Diagnosis Support Method) As described above, the pathological diagnosis support method S1 employs a configuration including an acquisition process S11 in which the acquisition unit 11 acquires a target image obtained by capturing, with a camera, an image of specimen cells to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination process S12 in which the determination unit 12 determines whether or not correction of the parameter acquired by the acquisition unit 11 is necessary; a notification process S14 in which the notification unit 13 notifies that correction of the parameter is necessary, if the determination unit 12 determines that correction is necessary; and a classification process S15 in which the classification unit 14 classifies the specimen cells as benign or malignant using the parameter, if the determination unit 12 determines that correction is not necessary.

[0037] Therefore, according to the pathological diagnosis support method S1, the same effects as those of the pathological diagnosis support device 1 described above can be obtained.

[0038] Furthermore, the pathology diagnosis support method S1 may employ a configuration including an acquisition process S11 in which an acquisition unit 11 acquires a target image obtained by a camera capturing an image of a specimen cell group to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination process S12 in which a judgment unit 12 judges whether or not correction of the parameter acquired by the acquisition unit 11 is necessary; a notification process S14 in which a notification unit 13 notifies that correction of the parameter is necessary if the judgment unit 12 judges that correction is necessary; and a classification process S15 in which a classification unit 14 uses the parameter to classify each specimen cell group as benign or malignant if the judgment unit 12 judges that correction is not necessary.

[0039] Therefore, the pathological diagnosis support method S1 has the effect of providing a technique for improving the accuracy of classifying each specimen cell group as benign or malignant.

[0040] Second Exemplary Embodiment A second exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same functions as those described in the above exemplary embodiment will be denoted by the same reference numerals, and their description will be omitted as appropriate. The scope of application of each technical means employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technical means employed in this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical hindrance occurs. Furthermore, each technical means shown in each drawing referenced to describe this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical hindrance occurs.

[0041] (Outline of Pathological Diagnosis Support Device 2) An outline of the pathological diagnosis support device 2 will be described with reference to Fig. 3. Fig. 3 is a diagram showing an example of a pathological diagnosis using the pathological diagnosis support device 2.

[0042] The pathological diagnosis support device 2 is a device that supports intraoperative rapid cytological diagnosis. As an example, as shown in Figure 3, the pathological diagnosis support device 2 acquires a target image op containing a specimen cell group as a subject from a camera CA. The pathological diagnosis support device 2 classifies each specimen cell group contained as a subject in the target image op as benign or malignant.

[0043] Note that the pathological diagnosis support apparatus 2 may be configured to acquire a target image op containing specimen cells as subjects from a camera CA, similar to the pathological diagnosis support apparatus 1 according to the exemplary embodiment described above. In this configuration, the pathological diagnosis support apparatus 2 performs benign / malignant classification of specimen cells contained as subjects in the target image op, determining whether the specimen cells are benign or malignant. Hereinafter, a specimen cell group may be a single specimen cell, and the benign / malignant classification performed on each specimen cell group may be the benign / malignant classification performed on a single specimen cell.

[0044] The camera CA is attached to the microscope MS as shown in Figure 3. The camera CA captures an image of a specimen cell group magnified by the microscope MS using the objective lens OL. The camera CA outputs a target image op obtained by capturing the image to the pathological diagnosis support device 2. The magnification of the objective lens OL of the microscope MS is not particularly limited, but an example is 40x.

[0045] Furthermore, the pathology diagnosis support device 2 standardizes the size of the subject included in the target image op and uses the standardized target image op to classify each specimen cell group as benign or malignant. Here, "standardization" refers to matching the size of the subject between multiple images. To standardize the size of the target image op, the pathology diagnosis support device 2 applies a parameter indicating an effective magnification to the target image op. In this specification, the effective magnification refers to the magnification of the size of the subject in the target image op relative to the actual size of the subject. The parameter indicating the effective magnification is not particularly limited as long as it indicates the effective magnification, but in one embodiment, it may be an effective magnification correction parameter (actually measured calibration value). The effective magnification correction parameter is a parameter indicating the length of the subject per pixel in the target image op, and may be, for example, a parameter generated by capturing an image of a specified micrometer.

[0046] In this configuration, the pathological diagnosis support device 2 determines whether parameter correction is necessary, and if it is determined that parameter correction is necessary, notifies the user that parameter correction is necessary. As an example, as shown in FIG. 3 , the pathological diagnosis support device 2 displays on the display DP that parameter correction is necessary. As another example, the pathological diagnosis support device 2 may notify the user that parameter correction is necessary by voice. On the other hand, if it determines that parameter correction is not necessary, the pathological diagnosis support device 2 applies the parameters to the target image op and classifies each of the specimen cell groups as benign or malignant.

[0047] (Configuration of pathological diagnosis support device 2) The configuration of the pathological diagnosis support device 2 will be described with reference to Fig. 4. Fig. 4 is a block diagram showing the configuration of the pathological diagnosis support device 2. As shown in Fig. 4, the pathological diagnosis support device 2 includes a control unit 20, a storage unit 31, an input unit 32, and an output unit 33.

[0048] The storage unit 31 stores data referenced by the control unit 20. As an example, the storage unit 31 stores parameters pm, test images tp, latent expressions le, and classification results cr, which will be described later, in association with each other. Examples of the storage unit 31 include, but are not limited to, flash memory, a hard disk drive (HDD), a solid state drive (SSD), or a combination thereof.

[0049] The input unit 32 accepts input of data and supplies the accepted data to the control unit 20. An example of the data accepted by the input unit 32 is a target image op captured by the camera CA.

[0050] The output unit 33 outputs data supplied from the control unit 20. As an example, the output unit 33 outputs an image to the display DP described above.

[0051] (Control Unit 20) The control unit 20 controls each component included in the pathological diagnosis support device 2. As shown in Fig. 4 , the control unit 20 also includes an acquisition unit 11, a determination unit 12, a notification unit 13, a classification unit 14, a correction unit 21, a calculation unit 22, and a selection unit 23. That is, the pathological diagnosis support device 2 includes the correction unit 21, the calculation unit 22, and the selection unit 23 in addition to the acquisition unit 11, the determination unit 12, the notification unit 13, and the classification unit 14 included in the pathological diagnosis support device 1. In this exemplary embodiment, the acquisition unit 11, the determination unit 12, the notification unit 13, the classification unit 14, the correction unit 21, the calculation unit 22, and the selection unit 23 are configured to realize an acquisition means, a determination means, a notification means, a classification means, a correction means, a calculation means, and a selection means, respectively.

[0052] The acquisition unit 11 acquires data and stores the acquired data in the storage unit 31.

[0053] As one example, the acquisition unit 11 acquires a target image op, which is an image of a specimen cell group to be classified as benign or malignant in intraoperative rapid cytology, captured by a microscope MS using a camera CA, and a parameter pm indicating an effective magnification, which is the magnification of the size of the subject in the target image op relative to the actual size of the subject. As another example, the acquisition unit 11 acquires the generation date and time when the parameter pm was generated. As yet another example, the acquisition unit 11 acquires the shipping date and time of the camera CA. As yet another example, the acquisition unit 11 acquires the detachment date and time when the camera CA was detached from the microscope MS.

[0054] An example of a method for the acquisition unit 11 to acquire data is to acquire data from a user via the input unit 32. Another example of a method for the acquisition unit 11 to acquire data is to acquire data from a camera CA via the input unit 32.

[0055] The determination unit 12 determines whether or not correction of the parameter pm is necessary, and stores the determination result in the storage unit 31. An example of the process executed by the determination unit 12 will be described later.

[0056] When the determination unit 12 determines that correction is necessary, the notification unit 13 notifies the user that correction of the parameter pm is necessary. As an example, the notification unit 13 displays information that correction of the parameter pm is necessary on the display DP via the output unit 33.

[0057] The classification unit 14 classifies the specimen cell group as benign or malignant. As an example, when the determination unit 12 determines that correction is unnecessary, the size of the target image op is standardized using the parameter pm, and the classification unit 14 performs benign or malignant classification on each specimen cell group included as a subject in the standardized target image op.

[0058] For example, if the standardized size indicates that the length per pixel is x [nm] and the parameter pm of the target image op indicates that the length per pixel is y [nm], the classification unit 14 standardizes the target image op by multiplying it by y / x.

[0059] An example of a method for the classification unit 14 to perform benign / malignant classification is a configuration using a machine-learned classification model. In this case, the classification unit 14 inputs an image including a specimen cell group as a subject, and inputs the target image op to which the parameter pm has been applied into a classification model that has been machine-learned to detect individual cells in the specimen cell group and classify each cell as benign or malignant, thereby performing benign / malignant classification for each specimen cell group included as a subject in the target image op. With this configuration, the classification unit 14 can input the standardized target image op to the classification model that has been machine-learned using the standardized image, thereby enabling favorable benign / malignant classification.

[0060] As described above, the classification unit 14 may be configured to classify the specimen cells included as a subject in the target image op as benign or malignant by inputting an image including the specimen cells as a subject, and inputting the target image op to which the parameters pm have been applied, into a classification model that has been machine-learned to classify the specimen cells as benign or malignant.

[0061] The correction unit 21 corrects the parameter pm and stores the corrected parameter pm in the storage unit 31. An example of the process executed by the correction unit 21 will be described later.

[0062] The calculation unit 22 calculates the similarity between each of the plurality of test images tp classified as benign or malignant by the classification unit 14 and the target image op. The calculation unit 22 stores the calculated similarity in the storage unit 31. The test image tp refers to an image that has been previously classified as benign or malignant for each of the specimen cell groups contained as subjects in the test image tp. The test image tp and the classification result cr may be stored in the storage unit 31 in association with each other, as shown in FIG. 4. An example of the processing performed by the calculation unit 22 will be described later.

[0063] The selection unit 23 selects a predetermined number of test images tp having high similarity calculated by the calculation unit 22. The selection unit 23 stores the selection results in the storage unit 31. The predetermined number is not particularly limited, but when there are N test images tp, it is, for example, N / 10.

[0064] (Pathological diagnosis support method S2) The flow of pathological diagnosis support method S2 will be described with reference to Fig. 5. Fig. 5 is a flow chart showing the flow of pathological diagnosis support method S2.

[0065] (Step S21) In step S21, the acquisition unit 11 acquires a target image op, which is an image of a specimen cell group to be classified as benign or malignant in intraoperative rapid cytology, captured by a microscope MS using a camera CA, and a parameter pm, which indicates an effective magnification, which is the magnification of the size of the subject in the target image op relative to the actual size of the subject. The acquisition unit 11 stores the acquired target image op and parameter pm in the storage unit 31.

[0066] (Step S22) In step S22, the determination unit 12 determines whether or not it is necessary to correct the parameter pm acquired by the acquisition unit 11. The determination unit 12 stores the determination result in the storage unit 31.

[0067] (Step S23) In step S23, the notification unit 13 and the classification unit 14 determine, based on the determination result stored in the storage unit 31, whether the determination unit 12 has determined that correction of the parameter pm is necessary.

[0068] (Step S24) If the determining unit 12 determines in step S23 that correction is necessary (step S23: YES), the notifying unit 13 notifies in step S24 that correction of the parameter pm is necessary.

[0069] (Step S25) In step S25, the correction unit 21 corrects the parameter pm. That is, the correction unit 21 corrects the parameter pm when the determination unit 12 determines that correction is necessary. The correction unit 21 stores the corrected parameter pm in the storage unit 31.

[0070] (Step S26) After step S25 is executed, or if the judgment unit 12 judges in step S23 that correction is not necessary (step S23: NO), in step S26, the classification unit 14 uses the parameters pm stored in the memory unit 31 to classify each of the specimen cell groups contained as subjects in the target image op into benign or malignant classes.

[0071] That is, when the classifying unit 14 executes step S26 after executing step S25, it performs benign / malignant classification for each of the specimen cell groups included as subjects in the target image op using the parameters pm corrected by the correcting unit 21. With this configuration, when it is determined that correction of the parameters pm is necessary, the classifying unit 14 performs benign / malignant classification for each of the specimen cell groups using the corrected parameters pm, thereby improving the accuracy of the benign / malignant classification.

[0072] (Example 1 of Processing Executed by Determination Unit 12) An example of processing executed by the determination unit 12 in step S22 will be described.

[0073] When the acquisition unit 11 acquires the creation date and time when the parameter pm stored in the storage unit 31 was created, the determination unit 12 determines whether or not correction of the parameter pm is necessary based on the creation date and time.

[0074] For example, if the acquisition unit 11 acquires the shipping date and time of the camera CA and the generation date and time at which the parameter pm was generated is the shipping date and time of the camera CA, the determination unit 12 determines that the parameter pm needs to be corrected. In this case, for example, the notification unit 13 notifies the user with a message saying, "Please photograph the micrometer and perform effective magnification correction."

[0075] In this way, when the correction unit 21 performs correction using an image of the micrometer, the notification unit 13 may notify the user in step S24 that the parameter pm needs to be corrected, including a notification that the micrometer will be placed in the field of view of the microscope MS (that the micrometer will be photographed). With this configuration, the pathological diagnosis support device 2 can preferably correct the parameter pm.

[0076] With this configuration, the pathological diagnosis support device 2 can correct the parameters pm according to the environment in which the camera CA is installed on the microscope MS.

[0077] As another example, if the acquisition unit 11 acquires the date and time of detachment of the camera CA and the generation date and time of the parameter pm is earlier than the date and time of detachment of the camera CA, the determination unit 12 determines that correction of the parameter pm is necessary. In this case, the notification unit 13 notifies, as an example, a message saying, "Effective magnification correction has not been performed since the camera was detached. Please re-photograph the micrometer and perform effective magnification correction."

[0078] With this configuration, the pathological diagnosis support device 2 can standardize the target image op even if the effective magnification changes before and after the camera CA is detached.

[0079] As yet another example, if the parameter pm was generated before a predetermined date and time, the determination unit 12 determines that the parameter pm needs to be corrected. In other words, if the parameter pm has not been updated for a predetermined period of time, the determination unit 12 determines that the parameter pm needs to be corrected. The predetermined period is not particularly limited, but an example is 30 days. In this case, the notification unit 13 notifies, as an example, the message, "The parameter for correcting the effective magnification has not been updated for a certain period of time. The last update date and time was XXXX year, Y month, Z day, t hour, u minute. If the microscope camera has been detached since the last update date, please re-photograph the micrometer and perform effective magnification correction."

[0080] With this configuration, the pathological diagnosis support device 2 can standardize the target image op even if the effective magnification changes over time.

[0081] In this way, the determination unit 12 determines whether or not correction of the parameter pm is necessary based on the generation date and time. Therefore, the determination unit 12 can correct the parameter pm when the effective magnification is changed by installing the camera CA, when the effective magnification is changed by attaching or detaching the camera CA, when the effective magnification is changed over time, etc.

[0082] (Second Example of Processing Executed by Determination Unit 12) Another example of processing executed by the determination unit 12 in step S22 will be described with reference to Fig. 6. Fig. 6 is a flowchart showing the flow of processing executed by the determination unit 12.

[0083] (Step S221) In step S221, the calculation unit 22 calculates the similarity between each of the plurality of test images tp stored in the storage unit 31 and the target image op.

[0084] An example of the process in step S221 in which the calculation unit 22 calculates the similarity will be described with reference to Fig. 7. Fig. 7 is a diagram showing an example of the process in which the calculation unit 22 calculates the similarity.

[0085] The calculation unit 22 calculates the similarity using, for example, an autoencoder AE. As shown in the upper part of Fig. 7, the autoencoder AE includes an encoder (Encoder) that has been machine-learned to receive an image as input, compress the dimensions of the image, and output a latent representation that represents an important feature, and a decoder (Decoder) that has been machine-learned to receive the latent representation as input and output an image decoded to its original dimensions. As an example, the latent representation is represented by a vector (e.g., 64 vectors).

[0086] The calculation unit 22 inputs a target image op to the autoencoder and obtains a latent representation le of the target image op. Similarly, the calculation unit 22 inputs a test image tp to the autoencoder and obtains a latent representation le of the test image tp.

[0087] The calculation unit 22 may obtain the latent representations le of the multiple test images tp before step S221 is executed. In this case, the calculation unit 22 may associate each test image tp with the latent representation le and store them in the storage unit 31, as shown in FIG.

[0088] Then, the calculation unit 22 calculates the similarity based on the difference between the latent expression l e of the target image op and the latent expression l e of the test image tp. As an example, the calculation unit 22 calculates the cosine similarity as the similarity between the target image op and each test image tp using the following formula (1):

[0089] With this configuration, the calculation unit 22 can preferably calculate the similarity between the target image op and the test image tp.

[0090] (Step S222) In step S222, the selection unit 23 selects a predetermined number of test images tp having high similarity calculated by the calculation unit 22.

[0091] (Step S223) In step S223, the classification unit 14 performs benign / malignant classification for each of the specimen cell groups included as subjects in the target image op.

[0092] (Step S224) In step S224, the judgment unit 12 judges whether correction of the parameter pm is necessary based on the difference between the classification results cr of each of the predetermined number of test images tp selected in step S222 and the classification result cr of the specimen cell group contained as a subject in the target image op in step S223.

[0093] As an example, if the classification result cr is the result of classification performed by the classification unit 14 using a classification model and is the probability that each individual sample cell group is malignant, and if the mean square error between the probability that the classification result cr is obtained for a predetermined number of test images tp and the probability that the classification result cr is obtained for the target image op is greater than or equal to a threshold, the judgment unit 12 judges that correction of the parameter pm is necessary.

[0094] In other words, if there is a large difference between the classification result cr of the test image tp similar to the target image op and the classification result cr of the target image op, the determination unit 12 determines that correction of the parameter pm is necessary.

[0095] In this way, the pathology diagnosis support device 2 determines whether correction of the parameter pm is necessary based on the difference between the classification result cr of the test image tp similar to the target image op and the classification result cr of the target image op. Specifically, if there is a large difference between the classification result cr of the test image tp similar to the target image op and the classification result cr of the target image op, the classification results cr will be significantly different despite the similarity between the target image op and the test image tp. In such a case, the effective magnifications at which the target image op and the test image tp were captured may be different. In such a case, it can be determined that correction of the parameter pm is necessary.

[0096] The determination unit 12 may be configured to combine the configuration of this example with a configuration that performs a determination based on the generation date and time when the parameter pm was generated. For example, the determination unit 12 may be configured to execute the process shown in FIG. 6 when the determination result based on the generation date and time indicates that correction of the parameter pm is unnecessary.

[0097] (Example of Processing Executed by Correction Unit 21) An example of processing executed by correction unit 21 in step S25 will be described with reference to Fig. 8 and Fig. 9. Fig. 8 is a flow chart showing the flow of processing executed by correction unit 21. Fig. 9 is an example of an image processed by correction unit 21.

[0098] (Step S251) In step S251, the acquisition unit 11 acquires an image including a micrometer as a subject. The correction unit 21 acquires the image of the micrometer acquired by the acquisition unit 11. For example, as shown in Fig. 9 , the correction unit 21 acquires an image Pic1 including a micrometer having a length of 1 mm as a subject.

[0099] (Step S252) In step S252, the correction unit 21 binarizes the image of the micrometer. For example, the correction unit 21 generates an image Pic2 by binarizing the image of the micrometer, as shown in FIG.

[0100] (Step S253) In step S253, the correction unit 21 performs expansion processing on the image generated in step S252. For example, the correction unit 21 performs expansion processing on the binarized image Pic2 to generate an image Pic3, as shown in FIG.

[0101] (Step S254) In step S254, the correction unit 21 extracts the object with the largest area from the image that has been subjected to the expansion process in step S253. For example, as shown in Fig. 9, the correction unit 21 generates an image Pic4 by maximizing the micrometer from the image Pic3 that has been subjected to the expansion process.

[0102] (Step S255) In step S255, the correction unit 21 performs a contraction process on the image generated in step S254. For example, as shown in FIG. 9, the correction unit 21 performs a contraction process on the image Pic4 in which the micrometer is maximized to generate an image Pic5.

[0103] (Step S256) In step S256, the correction unit 21 corrects the parameter pm.

[0104] For example, as shown in Fig. 9, the correction unit 21 detects points P1, P2, P3, and P4, which are the four ends of the micrometer, in the image Pic5 that has been subjected to the contraction process. In other words, it detects points P1, P2, P3, and P4, which are the farthest points from the center of the micrometer. Next, the correction unit 21 calculates the corrected parameter (length per pixel [nm]) using the following equation (2): Corrected parameter = 1.0 x 10 ６ / Ave(d_P12+d_P34) (2) d_P12: number of pixels from point P1 to point P2 d_P34: number of pixels from point P1 to point P2 (Effect of Pathological Diagnosis Support Apparatus 2) In this way, when the parameters pm to be applied to the target image op need to be corrected, the pathological diagnosis support apparatus 2 notifies the user that the parameters pm need to be corrected, and then corrects the parameters pm. The pathological diagnosis support apparatus 2 then performs a pathological diagnosis on each of the specimen cell groups using the corrected parameters pm.

[0105] For example, if multiple target images op are captured at different dates and times, if multiple target images op are captured using different microscopes MS, or if multiple target images op are captured using different cameras CA, the effective magnifications of the target images op may be different. In such cases, the pathology diagnosis support device 2 can standardize the multiple target images op to make subjects of the same size the same size across the multiple target images op.

[0106] Therefore, the pathological diagnosis support device 2 can improve the accuracy of classifying each specimen cell group into benign or malignant in pathological diagnosis.

[0107] In particular, when using a classification model that has been machine-learned to detect individual cells in a specimen cell group and classify each cell as benign or malignant, if the effective magnification of the image used to train the classification model differs from the effective magnification of the target image op, the accuracy of the benign / malignant classification by the classification model decreases. As described above, the pathology diagnosis support device 2 can standardize the image used to train the classification model and the target image op. Therefore, even when using a classification model, the pathology diagnosis support device 2 can improve the accuracy of benign / malignant classification of individual specimen cell groups.

[0108] [Example of Software Implementation] Some or all of the functions of the pathological diagnosis support devices 1 and 2 (hereinafter also referred to as "each of the above devices") may be implemented by hardware such as an integrated circuit (IC chip), or by software.

[0109] In the latter case, each of the above devices is realized by, for example, a computer that executes instructions of a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as computer C) is shown in Figure 10. Figure 10 is a block diagram showing the hardware configuration of computer C that functions as each of the above devices.

[0110] The computer C includes at least one processor C1 and at least one memory C2. The memory C2 stores a program P for causing the computer C to function as each of the above-mentioned devices. In the computer C, the processor C1 reads and executes the program P from the memory C2, thereby realizing the functions of each of the above-mentioned devices.

[0111] The processor C1 may be, for example, a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), a micro processing unit (MPU), a floating point number processing unit (FPU), a physics processing unit (PPU), a tensor processing unit (TPU), a quantum processor, a microcontroller, or a combination thereof. The memory C2 may be, for example, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or a combination thereof.

[0112] The computer C may further include a RAM (Random Access Memory) for expanding the program P during execution and for temporarily storing various data. The computer C may also include a communication interface for transmitting and receiving data to and from other devices. The computer C may also include an input / output interface for connecting input / output devices such as a keyboard, a mouse, a display, and a printer.

[0113] The program P can also be recorded on a non-transitory, tangible recording medium M that can be read by the computer C. Such a recording medium M can be, for example, a tape, a disk, a card, a semiconductor memory, or a programmable logic circuit. The computer C can acquire the program P via such a recording medium M. The program P can also be transmitted via a transmission medium. Such a transmission medium can be, for example, a communication network or broadcast waves. The computer C can also acquire the program P via such a transmission medium.

[0114] [Appendix A] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0115] (Appendix A1) A pathology diagnosis support device comprising: an acquisition means for acquiring a target image obtained by a camera capturing an image of specimen cells to be classified as benign or malignant under a microscope in intraoperative rapid cytology, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination means for determining whether correction of the parameter is necessary; a notification means for notifying that correction of the parameter is necessary, if the determination means determines that correction is necessary; and a classification means for classifying the specimen cells as benign or malignant using the parameter, if the determination means determines that correction is not necessary.

[0116] (Appendix A2) The pathology diagnosis support device according to Appendix A1, further comprising a correction means for correcting the parameters, wherein when the determination means determines that correction is necessary, the correction means corrects the parameters, and the classification means classifies the specimen cells as benign or malignant using the parameters corrected by the correction means.

[0117] (Supplementary Note A3) The pathology diagnosis support device according to Supplementary Note A1 or A2, wherein the notification means includes a notification that a micrometer should be installed in the field of view of the microscope in the notification that the parameter needs to be corrected.

[0118] (Appendix A4) The pathology diagnosis support device according to any one of Appendices A1 to A3, wherein the classification means classifies specimen cells included as a subject in a target image as benign or malignant by inputting the target image into a classification model that has been machine-learned to classify the specimen cells as benign or malignant.

[0119] (Supplementary Note A5) The pathology diagnosis support device according to any one of Supplementary Notes A1 to A4, wherein the acquisition means further acquires a generation date and time when the parameter was generated, and the determination means determines whether or not correction of the parameter is necessary based on the generation date and time.

[0120] (Supplementary Note A6) The pathology diagnosis support device according to Supplementary Note A5, wherein the acquisition means further acquires a shipping date and time of the camera, and the determination means determines that correction of the parameter is necessary when the generation date and time is the shipping date and time.

[0121] (Supplementary Note A7) The pathology diagnosis support device according to Supplementary Note A5, wherein the acquisition means further acquires a detachment date and time when the camera is detached, and the determination means determines that correction of the parameter is necessary if the generation date and time is earlier than the detachment date and time.

[0122] (Supplementary Note A8) The pathology diagnosis support device according to Supplementary Note A5, wherein the determining means determines that correction of the parameter is necessary when the generation date and time is before a predetermined date and time.

[0123] (Appendix A9) A pathology diagnosis support device according to any one of Appendices A1 to A4, further comprising: a calculation means for calculating a similarity between each of a plurality of test images classified as benign or malignant by the classification means and the target image; and a selection means for selecting a predetermined number of test images with high similarity, wherein the classification means classifies specimen cells contained as subjects in the target image as benign or malignant, and the determination means determines whether correction of the parameters is necessary based on a difference between the classification result of each of the plurality of test images selected by the selection means and the classification result of the specimen cells contained as subjects in the target image.

[0124] (Appendix A10) The pathology diagnosis support device according to Appendix A9, wherein the calculation means calculates the similarity based on a difference between a latent representation obtained by inputting the test image into an encoder that has been machine-learned to take an image as input and output a latent representation of the image, and a latent representation obtained by inputting the target image.

[0125] (Appendix A11) A pathology diagnosis support device comprising: an acquisition means for acquiring a target image obtained by a camera of an image of a specimen cell group to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination means for determining whether correction of the parameter is necessary; a notification means for notifying that correction of the parameter is necessary, if the determination means determines that correction is necessary; and a classification means for classifying each of the specimen cell groups as benign or malignant using the parameter, if the determination means determines that correction is not necessary.

[0126] [Appendix B] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0127] (Appendix B1) A pathology diagnosis support method including: an acquisition process in which at least one processor acquires a target image obtained by a camera of an image of specimen cells to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination process in which it is determined whether correction of the parameter is necessary; a notification process in which, if it is determined in the determination process that correction is necessary, it notifies that correction of the parameter is necessary; and a classification process in which, if it is determined in the determination process that correction is not necessary, it classifies the specimen cells as benign or malignant using the parameter.

[0128] (Appendix B2) The pathology diagnosis support method according to Appendix B1, wherein the at least one processor further includes a correction process for correcting the parameters, and when it is determined in the judgment process that correction is necessary, the at least one processor corrects the parameters in the correction process, and classifies the specimen cells as benign or malignant in the classification process using the parameters corrected in the correction process.

[0129] (Supplementary Note B3) The pathology diagnosis support method according to Supplementary Note B1 or B2, wherein in the notification process, the at least one processor includes, in the notification that the parameter needs to be corrected, a notification that a micrometer should be installed at a viewing angle of the microscope.

[0130] (Appendix B4) The pathology diagnosis support method according to any one of Appendices B1 to B3, wherein in the classification process, the at least one processor inputs an image including specimen cells as a subject into a classification model that has been machine-learned to classify the specimen cells as benign or malignant, thereby classifying the specimen cells included as a subject in the target image into benign or malignant.

[0131] (Supplementary Note B5) The pathology diagnosis support method according to any one of Supplementary Notes B1 to B4, wherein the at least one processor further acquires a generation date and time when the parameter was generated in the acquisition process, and determines whether or not correction of the parameter is necessary based on the generation date and time in the determination process.

[0132] (Appendix B6) The pathology diagnosis support method according to Appendix B5, wherein the at least one processor further acquires a shipping date and time of the camera in the acquisition process, and the at least one processor determines that correction of the parameter is necessary if the generation date and time is the shipping date and time in the determination process.

[0133] (Supplementary Note B7) The pathology diagnosis support method according to Supplementary Note B5, wherein the acquisition process further acquires a detachment date and time when the camera is detached, and the determination process determines that correction of the parameter is necessary if the generation date and time is earlier than the detachment date and time.

[0134] (Supplementary Note B8) The pathology diagnosis support method according to Supplementary Note B5, wherein in the determination process, the at least one processor determines that correction of the parameter is necessary if the generation date and time is before a predetermined date and time.

[0135] (Appendix B9) A pathology diagnosis support method according to any one of Appendices B1 to B4, wherein the at least one processor further includes a calculation process for calculating a similarity between each of a plurality of test images classified as benign or malignant in the classification process and the target image, and a selection process for selecting a predetermined number of test images with high similarity, wherein in the classification process, the at least one processor classifies specimen cells contained as subjects in the target image as benign or malignant, and in the determination process, the at least one processor determines whether correction of the parameters is necessary based on a difference between the classification result of each of the plurality of test images selected in the selection process and the classification result of the specimen cells contained as subjects in the target image.

[0136] (Appendix B10) The pathology diagnosis support method according to Appendix B9, wherein in the calculation process, the at least one processor calculates the similarity based on a difference between a latent representation obtained by inputting the test image into an encoder that has been machine-learned to take an image as input and output a latent representation of the image, and a latent representation obtained by inputting the target image.

[0137] (Appendix B11) A pathology diagnosis support method including: an acquisition process in which at least one processor acquires a target image, which is an image taken with a camera of a specimen cell group to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination process in which it is determined whether correction of the parameter is necessary; a notification process in which, if it is determined in the determination process that correction is necessary, it notifies that correction of the parameter is necessary; and a classification process in which, if it is determined in the determination process that correction is not necessary, it uses the parameter to classify each of the specimen cell group as benign or malignant.

[0138] [Appendix C] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0139] (Appendix C1) A pathology diagnosis support program that causes a computer to function as a pathology diagnosis support device, the pathology diagnosis support program causing the computer to function as: an acquisition means that acquires a target image obtained by taking a microscopic image of specimen cells to be classified as benign or malignant in intraoperative rapid cytology using a camera, and a parameter that indicates an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination means that determines whether correction of the parameter is necessary; a notification means that, when the determination means determines that correction is necessary, notifies the user that correction of the parameter is necessary; and a classification means that, when the determination means determines that correction is not necessary, uses the parameter to classify the specimen cells as benign or malignant.

[0140] (Appendix C2) A pathology diagnosis support program according to Appendix C1, further causing the computer to function as a correction means for correcting the parameters, wherein when the determination means determines that correction is necessary, the correction means corrects the parameters, and the classification means classifies the specimen cells as benign or malignant using the parameters corrected by the correction means.

[0141] (Supplementary Note C3) The pathological diagnosis support program according to Supplementary Note C1 or C2, wherein the notification means includes a message to install a micrometer at the viewing angle of the microscope in the message that the parameter needs to be corrected.

[0142] (Appendix C4) The pathology diagnosis support program according to any one of Appendices C1 to C3, wherein the classification means classifies specimen cells included as a subject in a target image as benign or malignant by inputting the target image into a classification model that has been machine-learned to classify the specimen cells as benign or malignant.

[0143] (Appendix C5) The pathology diagnosis support program according to any one of appendices C1 to C4, wherein the acquisition means further acquires a generation date and time when the parameter was generated, and the determination means determines whether or not correction of the parameter is necessary based on the generation date and time.

[0144] (Appendix C6) The pathology diagnosis support program according to Appendix C5, wherein the acquisition means further acquires a shipping date and time of the camera, and the determination means determines that correction of the parameter is necessary when the generation date and time is the shipping date and time.

[0145] (Appendix C7) The pathology diagnosis support program according to Appendix C5, wherein the acquisition means further acquires a detachment date and time when the camera is detached, and the determination means determines that correction of the parameter is necessary if the generation date and time is earlier than the detachment date and time.

[0146] (Supplementary Note C8) The pathology diagnosis support program according to Supplementary Note C5, wherein the determining means determines that correction of the parameter is necessary when the generation date and time is before a predetermined date and time.

[0147] (Appendix C9) A pathology diagnosis support program according to any one of appendices C1 to C4, further comprising: a calculation means for calculating the similarity between each of a plurality of test images classified as benign or malignant by the classification means and the target image; and a selection means for selecting a predetermined number of test images with high similarity, wherein the classification means classifies specimen cells contained as subjects in the target image as benign or malignant; and the determination means determines whether correction of the parameters is necessary based on the difference between the classification results of each of the plurality of test images selected by the selection means and the classification results of specimen cells contained as subjects in the target image.

[0148] (Appendix C10) The pathology diagnosis support program according to Appendix C9, wherein the calculation means calculates the similarity based on a difference between a latent representation obtained by inputting the test image into an encoder that has been machine-learned to take an image as input and output a latent representation of the image, and a latent representation obtained by inputting the target image.

[0149] (Appendix C11) A pathology diagnosis support program that causes a computer to function as a pathology diagnosis support device, the pathology diagnosis support program causing the computer to function as: an acquisition means that acquires a target image obtained by a camera of an image taken under a microscope of a specimen cell group to be classified as benign or malignant in intraoperative rapid cytology, and a parameter that indicates an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination means that determines whether correction of the parameter is necessary; a notification means that, when the determination means determines that correction is necessary, notifies the user that correction of the parameter is necessary; and, when the determination means determines that correction is not necessary, a classification means that uses the parameter to classify each of the specimen cell group as benign or malignant.

[0150] [Appendix D] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0151] (Appendix D1) A pathology diagnosis support device comprising at least one processor, the at least one processor executing: an acquisition process for acquiring a target image obtained by taking a microscopic image of specimen cells to be classified as benign or malignant in intraoperative rapid cytology with a camera, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination process for determining whether correction of the parameter is necessary; a notification process for notifying that correction of the parameter is necessary, if it is determined in the determination process that correction is necessary; and a classification process for classifying the specimen cells as benign or malignant using the parameter, if it is determined in the determination process that correction is not necessary.

[0152] (Appendix D2) The pathology diagnosis support device described in Appendix D1, wherein the at least one processor further performs a correction process to correct the parameters, and if it is determined in the judgment process that correction is necessary, corrects the parameters in the correction process, and classifies the specimen cells as benign or malignant using the parameters corrected by the correction process in the classification process.

[0153] (Supplementary Note D3) The pathology diagnosis support device according to Supplementary Note D1 or D2, wherein in the notification process, the at least one processor includes, in the notification that the parameter needs to be corrected, an instruction to install a micrometer at a viewing angle of the microscope.

[0154] (Appendix D4) The pathology diagnosis support device of any one of Appendices D1 to D3, wherein in the classification process, the at least one processor inputs an image including specimen cells as a subject into a classification model that has been machine-learned to classify the specimen cells as benign or malignant, thereby classifying the specimen cells included as a subject in the target image into benign or malignant.

[0155] (Supplementary Note D5) The pathology diagnosis support device according to any one of Supplementary Notes D1 to D4, wherein the at least one processor further acquires a generation date and time when the parameter was generated in the acquisition process, and determines whether or not correction of the parameter is necessary based on the generation date and time in the determination process.

[0156] (Appendix D6) The pathology diagnosis support device according to Appendix D5, wherein the at least one processor further acquires a shipping date and time of the camera in the acquisition process, and determines that correction of the parameter is necessary if the generation date and time is the shipping date and time in the determination process.

[0157] (Appendix D7) The pathology diagnosis support device according to Appendix D5, wherein the at least one processor further acquires a detachment date and time when the camera is detached in the acquisition process, and determines in the determination process that correction of the parameter is necessary if the generation date and time is earlier than the detachment date and time.

[0158] (Supplementary Note D8) The pathology diagnosis support device according to Supplementary Note D5, wherein in the determination process, the at least one processor determines that correction of the parameter is necessary if the generation date and time is before a predetermined date and time.

[0159] (Appendix D9) The pathology diagnosis support device of any one of Appendices D1 to D4, wherein the at least one processor further executes a calculation process to calculate a similarity between each of a plurality of test images classified as benign or malignant by the classification process and the target image, and a selection process to select a predetermined number of test images with high similarity, wherein the classification process classifies specimen cells contained as subjects in the target image as benign or malignant, and wherein the determination process determines whether correction of the parameters is necessary based on a difference between the classification result of each of the plurality of test images selected in the selection process and the classification result of the specimen cells contained as subjects in the target image.

[0160] (Appendix D10) The pathology diagnosis support device according to Appendix D9, wherein in the calculation process, the at least one processor calculates the similarity based on a difference between a latent representation obtained by inputting the test image into an encoder that has been machine-learned to take an image as input and output a latent representation of the image, and a latent representation obtained by inputting the target image.

[0161] [Appendix E] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0162] (Appendix E1) A non-transient recording medium having recorded thereon a pathology diagnosis support program that causes a computer to function as a pathology diagnosis support device, the pathology diagnosis support program causing the computer to execute the following: an acquisition process that acquires a target image obtained by taking a microscopic image of specimen cells to be classified as benign or malignant in intraoperative rapid cytology using a camera, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a determination process that determines whether correction of the parameter is necessary; a notification process that notifies the user that correction of the parameter is necessary if it is determined in the determination process that correction is necessary; and a classification process that uses the parameter to classify the specimen cells as benign or malignant if it is determined in the determination process that correction is not necessary.

[0163] 1, 2 Pathological diagnosis support device 11 Acquisition unit 12 Determination unit 13 Notification unit 14 Classification unit 21 Correction unit 22 Calculation unit 23 Selection unit cr Classification result le Latent expression op Target image tp Test image

Claims

1. A pathology diagnosis support device comprising: an acquisition means for acquiring a target image obtained by capturing an image of specimen cells to be classified as benign or malignant under a microscope using a camera, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a judgment means for judging whether or not correction of the parameter is necessary; a notification means for notifying the user that correction of the parameter is necessary when the judgment means judges that correction is necessary; and a classification means for classifying the specimen cells as benign or malignant using the parameter when the judgment means judges that correction is not necessary.

2. A pathology diagnosis support device as described in claim 1, further comprising a correction means for correcting the parameters, wherein when the determination means determines that correction is necessary, the correction means corrects the parameters, and the classification means uses the parameters corrected by the correction means to classify the specimen cells as benign or malignant.

3. The pathology diagnosis support device according to claim 1 or 2, wherein the notification means includes a notification that a micrometer should be installed at the viewing angle of the microscope in the notification that the parameter needs to be corrected.

4. A pathology diagnosis support device as claimed in any one of claims 1 to 3, wherein the classification means classifies specimen cells contained as a subject in a target image as benign or malignant by inputting the target image into a classification model that has been machine-learned to classify the specimen cells as benign or malignant.

5. A pathology diagnosis support device as described in any one of claims 1 to 4, wherein the acquisition means further acquires a generation date and time when the parameter was generated, and the determination means determines whether or not correction of the parameter is necessary based on the generation date and time.

6. The pathology diagnosis support device according to claim 5, wherein the acquisition means further acquires a shipping date and time of the camera, and the determination means determines that correction of the parameter is necessary if the generation date and time is the shipping date and time.

7. The pathology diagnosis support device according to claim 5, wherein the acquisition means further acquires a date and time when the camera was detached, and the determination means determines that correction of the parameter is necessary if the generation date and time is earlier than the date and time of detachment.

8. The pathology diagnosis support device according to claim 5, wherein said determination means determines that correction of said parameters is necessary if said generation date and time is earlier than a predetermined date and time.

9. A pathology diagnosis support device as described in any one of claims 1 to 4, further comprising: a calculation means for calculating a similarity between each of a plurality of test images classified as benign or malignant by the classification means and the target image; and a selection means for selecting a predetermined number of test images with high similarity, wherein the classification means classifies specimen cells contained as subjects in the target image as benign or malignant, and the determination means determines whether or not correction of the parameters is necessary based on a difference between the classification result of each of the plurality of test images selected by the selection means and the classification result of specimen cells contained as subjects in the target image.

10. The pathology diagnosis support device according to claim 9, wherein the calculation means calculates the similarity based on a difference between a latent representation obtained by inputting the test images into an encoder that has been machine-learned to take an image as input and output a latent representation of the image, and a latent representation obtained by inputting the target image.

11. A pathology diagnosis support method including: an acquisition process in which at least one processor acquires a target image captured by a camera of an image of specimen cells to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a judgment process in which it is determined whether or not correction of the parameter is necessary; a notification process in which, if it is determined in the judgment process that correction is necessary, it notifies that correction of the parameter is necessary; and, if it is determined in the judgment process that correction is not necessary, a classification process in which the parameter is used to classify the specimen cells as benign or malignant.

12. A program that causes a computer to function as a pathology diagnosis support device, the program causing the computer to execute the following processes: an acquisition process that acquires a target image captured by a camera of an image of specimen cells to be classified as benign or malignant under a microscope, and a parameter indicating an effective magnification, which is the magnification of the size of the subject in the target image relative to the actual size of the subject; a judgment process that determines whether or not correction of the parameter is necessary; a notification process that, if it is determined in the judgment process that correction is necessary, notifies the user that correction of the parameter is necessary; and, if it is determined in the judgment process that correction is not necessary, a classification process that uses the parameter to classify the specimen cells as benign or malignant.

Images