WO2012105281A1 - 情報処理システム、情報処理方法、情報処理装置及びその制御方法とその制御プログラムを格納した記憶媒体 - Google Patents
情報処理システム、情報処理方法、情報処理装置及びその制御方法とその制御プログラムを格納した記憶媒体 Download PDFInfo
- Publication number
- WO2012105281A1 WO2012105281A1 PCT/JP2012/050248 JP2012050248W WO2012105281A1 WO 2012105281 A1 WO2012105281 A1 WO 2012105281A1 JP 2012050248 W JP2012050248 W JP 2012050248W WO 2012105281 A1 WO2012105281 A1 WO 2012105281A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- regions
- similarity
- information processing
- roi
- tissue specimen
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
- G06T7/0014—Biomedical image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
- G06V20/695—Preprocessing, e.g. image segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30024—Cell structures in vitro; Tissue sections in vitro
Definitions
- the present invention relates to an information processing technique for evaluating diagnosis based on a tissue specimen image of a living tissue.
- Patent Document 1 discloses a technique for automatically extracting and ranking a region of interest (ROI) that is important for diagnosis in order to support diagnosis from a tissue specimen image that is a pathological image. Is disclosed.
- ROI region of interest
- the above prior art compares the selection of the ROI by the pathologist for the same tissue specimen image with the selection of the ROI by the apparatus, or the selection of the ROI for the same tissue specimen image by a different pathologist. There is no mention of assessing the correctness of the choice. For this reason, it has not been possible to determine whether or not the ROI selected by the pathologist or device includes an area important for diagnosis.
- An object of the present invention is to provide a technique for solving the above-described problems.
- an apparatus provides: A first plurality of regions selected as a diagnosis target from a tissue specimen image obtained by imaging a biological tissue, a second plurality of regions selected as a diagnosis target from the tissue sample image, and the selected on the tissue sample image Input means for inputting the position information of each area, A computing means for computing the similarity between the first plurality of regions and the second plurality of regions based on a correlation considering a distance on the tissue specimen image between the selected regions; It is characterized by providing.
- the method according to the present invention comprises: A first plurality of regions selected as a diagnosis target from a tissue specimen image obtained by imaging a biological tissue, a second plurality of regions selected as a diagnosis target from the tissue sample image, and the selected on the tissue sample image An input step for inputting the position information of each area, A calculation step of calculating a similarity between the first plurality of regions and the second plurality of regions based on a correlation in consideration of a distance on the tissue specimen image between the selected regions; It is characterized by including.
- a storage medium provides: A first plurality of regions selected as a diagnosis target from a tissue specimen image obtained by imaging a biological tissue, a second plurality of regions selected as a diagnosis target from the tissue sample image, and the selected on the tissue sample image An input step for inputting the position information of each area, A calculation step of calculating a similarity between the first plurality of regions and the second plurality of regions based on a correlation in consideration of a distance on the tissue specimen image between the selected regions; A control program for causing a computer to execute is stored.
- an apparatus provides: Transmitting means for selecting a diagnosis target from a tissue specimen image obtained by imaging a biological tissue and transmitting the first plurality of regions; In response to the transmission of the first plurality of regions, the first plurality of regions and the second plurality of regions selected as diagnostic targets from the tissue sample image are used according to the feature amount of the tissue sample image in each region.
- Receiving means for receiving the weighted similarity calculated based on the distance between the selected regions and the second plurality of regions; Display means for displaying the similarity received by the receiving means and the first and second plurality of areas that can be identified; It is characterized by providing.
- the method according to the present invention comprises: A transmission step of transmitting a first plurality of regions selected as diagnostic targets from a tissue specimen image obtained by imaging a biological tissue; In response to the transmission of the first plurality of regions, the first plurality of regions and the second plurality of regions selected as diagnostic targets from the tissue sample image are used according to the feature amount of the tissue sample image in each region. Receiving a weighted similarity calculated based on the distance between each selected region and the second plurality of regions; A display step for displaying the similarity received by the receiving step and the first and second plurality of regions that can be identified; It is characterized by including.
- a storage medium provides: A transmission step of transmitting a first plurality of regions selected as diagnostic targets from a tissue specimen image obtained by imaging a biological tissue; In response to the transmission of the first plurality of regions, the first plurality of regions and the second plurality of regions selected as diagnostic targets from the tissue sample image are used according to the feature amount of the tissue sample image in each region. Receiving a weighted similarity calculated based on the distance between each selected region and the second plurality of regions; A display step for displaying the similarity received by the receiving step and the first and second plurality of regions that can be identified; A control program for causing a computer to execute is stored.
- a system provides: An information processing system that supports diagnosis based on a tissue specimen image obtained by imaging a biological tissue, A first plurality of regions selected as diagnostic targets from the tissue specimen image; a second plurality of regions selected as diagnostic targets from the tissue specimen image; and the selected regions on the tissue specimen image.
- Input means for inputting position information;
- a computing means for computing the similarity between the first plurality of regions and the second plurality of regions based on a correlation considering a distance on the tissue specimen image between the selected regions;
- Display means for displaying the similarity calculated by the calculation means and the first and second plurality of areas that can be identified; It is characterized by providing.
- the method according to the present invention comprises: An information processing method for supporting diagnosis based on a tissue specimen image obtained by photographing a biological tissue, A first plurality of regions selected as diagnostic targets from the tissue specimen image; a second plurality of regions selected as diagnostic targets from the tissue specimen image; and the selected regions on the tissue specimen image.
- An input step for inputting location information A calculation step of calculating a similarity between the first plurality of regions and the second plurality of regions based on a correlation in consideration of a distance on the tissue specimen image between the selected regions;
- the present invention it is possible to evaluate whether or not the ROI selected by the pathologist or device includes an important region of interest.
- the information processing apparatus 100 includes an input unit 110 and a calculation unit 120.
- the input unit 110 includes a first plurality of regions to be diagnosed selected from a tissue specimen image obtained by imaging a biological tissue, a second plurality of regions to be diagnosed selected from a tissue specimen image, and a tissue sample image And the position information of each of the selected areas.
- the computing unit 120 computes the similarity between the first plurality of regions and the second plurality of regions based on the correlation considering the distance on the tissue specimen image between the selected regions.
- the present embodiment it is possible to evaluate whether or not the ROI selected by the pathologist or device includes an important region of interest.
- the similarity to the selection result of other ROIs for the same tissue region is calculated to evaluate whether the ROI has been successfully selected.
- the selection of another ROI is automatically performed by the pathological image analysis support apparatus of the analysis center. Therefore, according to this embodiment, the analysis center can quantitatively evaluate the selection of a plurality of ROIs transmitted from the pathological medical terminal.
- FIG. 2 is a block diagram illustrating a configuration of an information processing system 200 including an analysis center 210 that is an information processing apparatus according to the present embodiment.
- An information processing system 200 in FIG. 2 includes an analysis center 210 that is an information processing apparatus that supports pathological image diagnosis, and a plurality of pathological medical terminals 220 that receive support for pathological image diagnosis connected to the analysis center 210 via a network 230. With.
- the analysis center 210 has a communication control unit 215 for communicating with the network 230.
- the ROI receiving unit 211 receives the ROI information from the pathological medical terminal 220 received by the communication control unit 215 together with the ROI analysis priority.
- the priority order of ROI from the pathologist terminal 220 is hereinafter referred to as the first priority order.
- a tissue specimen image storage unit 212 that stores the tissue specimen image received by the communication control unit 215 is provided. Based on the tissue specimen image stored in the tissue specimen image storage unit 212, an ROI setting unit 213 is provided for setting ROI and ROI analysis priority.
- the priority order of ROI set by the analysis center 210 is hereinafter referred to as the second priority order.
- the tissue specimen image stored in the tissue specimen image storage unit 212 includes an ROI feature amount analysis unit 214 that analyzes ROI feature amounts indicated by the ROI reception unit 211 and the ROI setting unit 213.
- the similarity calculation unit 216 uses the ROI and first priority of the ROI receiving unit 211, the ROI and second priority of the ROI setting unit 213, and the feature amount of each ROI analyzed by the ROI feature amount analysis unit 214. Thus, the similarity (correlation) between the ROI of the ROI receiving unit 211 and the ROI of the ROI setting unit 213 is calculated.
- the similarity calculated by the similarity calculation unit 216 is transmitted together with the ROI of the ROI setting unit 213 by the similarity / ROI transmission unit 217 via the communication control unit 215 to the request source that is the input source of the diagnosis support. Further, since the similarity calculated by the similarity calculation unit 216 is referred to for subsequent diagnosis support, it is stored in the similarity storage DB 218 so as to be searchable in the request source unit.
- Each pathological terminal 220 includes a control unit 221 that controls the operation of the pathological terminal 220 and communication with the analysis center 210.
- a scanner 222 that reads a pathological slide obtained by photographing a stained biological tissue with a resolution capable of diagnosis is provided.
- a display 223 for displaying the tissue specimen image read by the scanner 222 is provided.
- FIG. 2 does not show a keyboard or pointing device for data input or operation instruction, but it is assumed that necessary input / output devices are connected.
- FIG. 3 is a sequence diagram showing an operation procedure 300 of the information processing system including the information processing apparatus according to the present embodiment.
- FIG. 3 operations from reading a pathological slide by the scanner 222 of the pathological medical terminal 220 to displaying the diagnosis support information on the screen will be described.
- the pathological medical terminal 220 reads a tissue specimen image from the pathological slide by the scanner 222.
- the read tissue specimen image is displayed on the display 223.
- a tissue region used for diagnosis is selected from a plurality of tissue regions in the tissue specimen image.
- an ROI that requests analysis to the analysis center 210 for diagnosis support is selected from the selected tissue region (see FIG. 4).
- the selection of the tissue region and the selection of the ROI may be instructed by the pathologist from the tissue specimen image on the display 223 screen, or may be determined by existing automatic ROI setting software.
- the pathological medical terminal 220 transmits the tissue specimen image, the selected ROI, and the priority of the ROI to the analysis center 210.
- At least the terminal ID of the pathologist terminal 220, the image ID for specifying the image, and the doctor ID for specifying the pathologist are attached to the tissue specimen image to be transmitted in order to transmit the processing result.
- the site of the taken biological tissue eg, stomach, lung, breast, prostate
- staining method eg, HE method, IHC method, FISH method
- gender and age information, address and nationality information may be added for analysis or information storage in a database (hereinafter referred to as DB) or analysis within a range in which the patient's personal information does not leak.
- DB database
- the analysis center 210 In response to the reception of the tissue specimen image, the analysis center 210 storing the tissue specimen image performs a tissue structure analysis using the tissue structure analysis DB previously learned and registered based on the tissue specimen image in step S307.
- the ROI unique to the center 210 is set and the priority of the ROI is set.
- the analysis center 210 analyzes each feature quantity of the ROI transmitted from the pathological medical terminal 220 and the ROI set by the analysis center 210 using the feature quantity analysis DB registered in advance. To do.
- step S311 the analysis center 210 determines the similarity (rank correlation coefficient) between the ROI transmitted from the pathological terminal 220 and the ROI set by the analysis center 210 based on the position information and feature amount of each ROI. Is calculated.
- step S313 the analysis center 210 transmits the analysis result and the calculated similarity to the pathological medical terminal 220. Further, the calculated similarity of the ROI setting is stored in the similarity storage DB 218 in association with the doctor ID and the image ID.
- the pathological medical terminal 220 receives the analysis result and the similarity, the pathological medical terminal 220 superimposes the analysis result and the similarity on the tissue specimen image read from the pathological slide in step S301 in step S315, and displays the result on the display 223 in step S317. (See FIG. 5).
- the pathologist refers to the analysis result displayed on the display 223 as support information, diagnoses the tissue specimen image, and can evaluate and learn his / her ROI setting based on the similarity.
- FIG. 4 is a diagram showing a screen 400 displayed on the display 223 of the pathological terminal 220 at the time of transmitting the ROI image from the pathological terminal according to the present embodiment.
- the screen 400 displays a plurality of selected ROIs 401 to 404 superimposed on the tissue region selected from the tissue specimen image.
- the priority order is indicated by a circle number.
- the tissue specimen image, the position information of the plurality of ROIs 401 to 404, and the priority order are transmitted to the analysis center 210 to obtain diagnosis support information.
- FIG. 4 shows a case where the ROI is rectangular, it may be another shape such as a circle or an ellipse, or may be a shape that matches the outline of the cell mass.
- the central position of the ROI is transmitted as the position information.
- reference numeral 405 in FIG. 4 denotes management information of the displayed tissue specimen image in the pathologist terminal 220 and information specifying the analysis center 210 that is a request destination for requesting diagnosis support. This includes the doctor ID of the pathologist who is the requester. Note that the information shown in 405 is an example, and the present invention is not limited to this.
- FIG. 5 is a diagram showing a screen 500 displayed on the display 223 of the pathologist terminal 220 when the analysis result and the similarity are received according to the present embodiment.
- the analysis results of the plurality of ROIs 401 to 404 in FIG. 4 are represented by display of feature amounts analyzed corresponding to each ROI having cancer cells.
- ROIs 501 and 504 indicate that there are no cancer cells because no feature value is displayed.
- the ROIs 502 and 503 are displayed as values of the average nucleus size ( ⁇ m 2 ), average deformity, and texture as feature amounts.
- ROIs 505 and 506 in FIG. 5 are ROIs set by the analysis center 210, which are different from the ROIs 501 and 504 transmitted from the pathological medical terminal 220.
- ROI 506 indicates an area without cancer cells, similar to ROI 504.
- the ROI 505 is a ROI having a characteristic amount displayed and having cancer cells.
- the solid circled numbers indicate the priorities of the ROIs 501 to 504 at the time of transmission, and the identifiable broken line circles indicate the priorities of the ROIs 503, 502, 505, and 506 set by the analysis center 210.
- 5 is management information of the displayed tissue specimen image in the pathologist terminal 220 and information specifying the analysis center 210 of the report source that has reported the analysis result of diagnosis support. 510 indicates that the similarity is 0.65. Note that the information shown in 510 is an example, and the present invention is not limited to this. The pathologist can recognize the evaluation of the setting of his / her ROI 401 to 404 by looking at the display.
- FIG. 6 is a block diagram showing a hardware configuration of the analysis center 210 according to the present embodiment. Although FIG. 6 shows a configuration with one device, it may be configured with a plurality of devices according to function.
- a CPU 610 is a processor for arithmetic control, and implements each functional component of the analysis center 210 by executing a program.
- the ROM 620 stores fixed data and programs such as initial data and programs.
- the communication control unit 215 controls communication with a plurality of pathological medical terminals 220 via the network 230. Such communication may be wired or wireless.
- the RAM 640 is a random access memory that the CPU 610 uses as a temporary storage work area. In the RAM 640, an area for storing data necessary for realizing the present embodiment is secured. Each region stores tissue specimen image data received from the pathologist terminal 220 and received data 641 including ROI position information. In addition, tissue specimen image data 642 received from the pathologist terminal 220 is stored. Further, similarity data 643 calculated by the analysis center 210 is stored. In addition, transmission data 644 including an analysis result and similarity to be transmitted to the pathologist terminal 220 is stored.
- the storage 650 is a mass storage device that stores a database, various parameters, and a program executed by the CPU 610 in a nonvolatile manner.
- the storage 650 stores the following data or programs necessary for realizing the present embodiment.
- a feature quantity analysis DB 651 used for performing ROI feature quantity analysis using tissue specimen image data is stored.
- a similarity accumulation DB 218 that accumulates the calculated similarity is stored (see FIG. 7).
- a similarity calculation algorithm 652 that is an algorithm including an arithmetic expression for calculating the similarity is stored.
- the feature amount analysis DB 651 is preferably updated by learning using image data received from the pathologist terminal 220, feedback of analysis results, statistical processing of analysis results, and the like.
- a pathological image diagnosis support program 653 for realizing a series of pathological image diagnosis support is stored as a program (see FIG. 8A).
- a feature amount analysis module 654 that performs feature amount analysis of ROI using tissue specimen image data using the feature amount analysis DB 651 constituting a part of the pathological image diagnosis support program 653 is stored.
- a similarity calculation module 655 that calculates the similarity using the similarity calculation algorithm 652 is stored.
- an analysis result transmission module 656 that transmits the analysis result and the similarity to the pathological medical terminal 220 as diagnosis support information is stored.
- FIG. 6 shows only data and programs essential for the present embodiment, and general-purpose data and programs such as OS are not shown.
- FIG. 7 is a block diagram showing the configuration of the similarity accumulation DB 218 according to this embodiment.
- Reference numeral 701 is a doctor ID for identifying the pathologist who requested the diagnosis support, and 702 is the request date and time.
- Reference numeral 703 denotes a tissue specimen image ID that identifies the received tissue specimen image data.
- Reference numerals 704 to 706 denote a sex / age for classifying the tissue specimen image data specified by the tissue specimen image ID 703, a region of the tissue specimen image to be analyzed, and a staining method of the biological tissue. This is information related to the analysis method.
- Reference numeral 707 denotes a calculated similarity that is a correlation value between the ROI received from the pathological medical terminal 220 and the ROI set in the analysis center 210.
- ROI information that is the basis for calculating each similarity 707 is stored in correspondence with each tissue specimen image received from the pathological medical terminal 220.
- the request setting information 712 includes information 712a indicating a tissue region in the tissue specimen image, an ROI centroid position address 712b indicating the centroid position of the ROI selected by the pathologist from the tissue region, and the ROI analyzed by the analysis center 210. And a feature amount 712c.
- the comparison setting information 713 includes information 713a indicating the set tissue region in the tissue specimen image, an ROI centroid position address 713b indicating the centroid position of the ROI selected from the tissue region by the analysis center 210, and the analysis center 210. Are included in the ROI feature quantity 713c analyzed. Note that the stored data of the ROI centroid position differs depending on the ROI shape. Similar ROI information 720 is stored as ROI information for which the similarity “0.74” in the third row is calculated. As described above, in this embodiment, the ROI barycentric position is set as the center of the ROI shape in consideration of the change in the ROI shape and the weighting of the pixel unit or cell unit in the ROI. However, it is not limited to the position of the center of gravity.
- FIG. 8A is a flowchart illustrating a processing procedure of the information processing apparatus according to the present embodiment. This flowchart is executed by the CPU 610 in FIG. 6 using the RAM 640, and implements each functional component of the analysis center 210 in FIG.
- step S801 reception of an analysis image and ROI from the pathological medical terminal 220 is awaited. If there is image reception, the process advances to step S803 to store the received tissue specimen image, ROI, and ROI priority. In step S805, the analysis center 210 sets ROI and ROI priority from the received tissue specimen image. Next, the process proceeds to step S807, and feature amount analysis of each ROI corresponding to the site, staining method, sex / age, etc. is performed.
- the feature amount analysis performed here is, for example, analysis of the size and shape of the cell nucleus in the case of HE-stained stomach tissue.
- step S809 the similarity (rank correlation coefficient) between the ROI from the pathological medical terminal 220 and the ROI set in the analysis center 210 is calculated.
- step S811 the feature amount analysis result and the similarity are transmitted with the image ID from the transmission source attached to the pathological medical terminal 220 of the transmission source.
- step S813 the similarity calculated in the similarity storage DB 218 is recorded in association with the doctor ID.
- the feature quantities analyzed in step S807 in FIG. 8A include the following feature quantities.
- the following example is an example in the case of HE dyeing
- Signet ring a signet ring cell
- feature values F1 to F7 are important features in almost any cancer. It has become. F1) the size of the nucleus, F2) the major and minor axis of the nucleus, F3) Circularity (maximum value 1 if close to a circle, smaller value if the degree of deviation from the circle is larger), F4) Texture, F5) Color (RGB), F6) Color (HSV), F7) Ductal area, There is.
- tissue structure analysis is performed with low-magnification image data
- more detailed feature analysis is performed with high-magnification image data.
- the feature quantities having the same name used in the tissue structure analysis and the feature quantity analysis are not the same because the resolutions of the images are different.
- the size of the nucleus is roughly analyzed by extracting a region stained with hematoxin and classifying it into a large nucleus and a small nucleus based on the pixel size.
- the outline of the nucleus is accurately extracted, and the size (or lightness of the circle, etc.) is calculated based on the outline.
- a gland duct area is extracted by analyzing the low-magnification image data of the tissue specimen image to generate a duct mask, and the mask information is As it is, it is passed to the high magnification image data analysis module. Based on this information, the high-magnification image data analysis module checks whether the gland duct contains the nucleus to be analyzed. If it is contained in the gland duct, It is also performed so that it is not judged.
- a rank correlation coefficient Rl for calculating the similarity of ROI selection depending on the feature in the pathological image is introduced.
- This rank correlation coefficient is obtained by linearly summing 1 minus the number of distances between the centroids of the ROI while weighting with a weighting coefficient that is a function of the feature quantity, and is represented by the following equation.
- N is the total number of classifications (target area: ROI)
- dn is the distance between the area centroids.
- wn (f1,..., fM) is a weighting coefficient determined by the amounts included in the two areas to be compared with the feature amounts f1,..., fM, and the maximum value is “1”.
- the weighting factors wn (f1,..., FM)
- first, one of the feature values is set to (f), the area selected by the first ROI selection to be compared is set to A1, and the second The area selected by ROI selection is A2.
- the smaller feature quantity f included in each is used as a reference quantity, and when the quantity exceeds a certain value “f0”, it is set to “1”, and when it is less than “f0”, the feature quantity of the area A1 A smaller value of “f1” and the feature amount “f2” of the area A2 is divided by “f0”.
- “f0”, “f1”, “f2”, and the like are specific numerical values, and are different from the above-described types of feature values f1, f2, and the like.
- FIG. 8B is a schematic diagram for schematically explaining the correlation value calculation between the two regions A1 and A2.
- the distance between the ROIs is dn, and the respective feature values are “f1” and “f2”.
- the calculated value is wn (f) ⁇ (1 ⁇ dn) shown in the figure.
- the above [Equation 1] is obtained by linearly summing such operations for each ROI.
- the maximum value of the rank correlation Rl becomes “1” (when the regions match and the feature value f equal to or greater than the fixed value “f0” is included in both selected regions). Since this correlation also depends on the feature quantity of the region, the correlation value will not increase even if the location is the same unless the pathological feature quantity is obtained in both. That is, there is an effect of preventing high correlation due to accidental coincidence with a weak pathological basis.
- FIG. 9 is a block diagram illustrating a hardware configuration of the pathological medical terminal 220 according to the present embodiment.
- the pathological medical terminal 220 includes a control unit 221, a scanner 222, and a display 223 as a basic configuration.
- a CPU 910 is a processor for arithmetic control, and realizes a control unit of the pathologist terminal 220 by executing a program.
- the ROM 920 stores fixed data and programs such as initial data and programs.
- the communication control unit 930 controls communication with the analysis center 210 via the network 230. Such communication may be wired or wireless.
- the RAM 940 is a random access memory that the CPU 910 uses as a work area for temporary storage. In the RAM 940, an area for storing data necessary for realizing the present embodiment is secured. In each region, tissue specimen image read data 941 read from the pathological slide by the scanner 222 is stored. Further, the image identification table 1042 for managing the image data of the tissue specimen image to be transmitted to the analysis center 210 and the notified similarity (rank correlation coefficient) and specifying the patient, the region, the ROI, and the like is stored. (See FIG. 10). Also, transmission / reception data 943 transmitted / received to / from the analysis center 210 is stored (see FIG. 10). Further, display data 944 to be displayed on the display 223 of the pathologist terminal 220 is stored.
- the storage 950 is a mass storage device that stores a database, various parameters, and a program executed by the CPU 910 in a nonvolatile manner.
- the storage 950 stores the following data or programs necessary for realizing the present embodiment.
- a tissue specimen image DB 951 read by the scanner 222 stored locally by a pathologist is stored.
- a patient history DB 952 that stores a diagnosis history corresponding to a patient is stored.
- a pathological image diagnosis processing program 953 including a process for requesting the analysis center 210 to provide pathological image diagnosis support is stored as a program (see FIG. 11).
- a transmission / reception control module 954 that controls data communication with the analysis center 210 that constitutes a part of the pathological image diagnosis processing program 953 is stored.
- an analysis result display module 955 that superimposes and displays the analysis result received from the analysis center 210 on the tissue specimen image is stored.
- the input interface 960 is an interface for inputting control signals and data necessary for control by the CPU 910.
- image data of a tissue specimen image obtained by reading a pathological slide from the scanner 222 is input. Note that other keyboards and pointing devices are not shown.
- the output interface 970 is an interface that outputs a control signal and data to the device under the control of the CPU 910.
- the tissue sample image, the diagnosis support request information to the analysis center 210, or the analysis result and similarity transmitted from the analysis center 210 are output to the display 223.
- FIG. 9 shows only data and programs essential to the present embodiment, and general-purpose data and programs such as OS are not shown.
- FIG. 10 is a diagram showing a configuration of the image identification table 942 and transmission / reception data 943 according to the present embodiment.
- 1001 is a patient ID for identifying a patient
- 1002 is a photographing date and time of a tissue specimen image
- 1003 is a tissue specimen image ID for identifying a tissue specimen image.
- Reference numeral 1004 denotes the sex / age of the patient
- reference numeral 1005 denotes a region of the tissue specimen image to be analyzed
- reference numeral 1006 denotes a staining method of the living tissue
- Reference numeral 1007 denotes a calculated similarity that is a correlation value between the ROI received from the pathological medical terminal 220 and the ROI set in the analysis center 210.
- ROI information that is the basis of calculation of each similarity 1007 is stored in correspondence with each tissue specimen image that has been read by a scanner and requested diagnosis support.
- ROI information 1010 in which the similarity “0.65” in the second row is calculated ROI priority 1011, request setting information 1012 transmitted to the analysis center 210, comparison setting information 1013 set by the analysis center 210, Is stored.
- the request setting information 1012 includes information 1012a indicating a tissue region in the tissue specimen image, an ROI centroid position address 1012b indicating the centroid position of the ROI selected by the pathologist from the tissue region, and the ROI analyzed by the analysis center 210. And a feature amount 1012c.
- the comparison setting information 1013 includes information 1013a indicating the set tissue region in the tissue specimen image, an ROI centroid position address 1013b indicating the centroid position of the ROI selected from the tissue region by the analysis center 210, and the analysis center 210. Are included in the ROI feature quantity 1013c analyzed. Note that the stored data of the ROI centroid position differs depending on the ROI shape. Similar ROI information 1020 is stored as ROI information for which the similarity “0.74” in the third row is calculated. As described above, in this embodiment, the ROI barycentric position is set as the center of the ROI shape in consideration of the change in the ROI shape and the weighting of the pixel unit or cell unit in the ROI. However, it is not limited to the position of the center of gravity.
- the partial information that is highly necessary for the above information may be stored in the pathological medical terminal 220 as a DB, but since it is stored in the analysis center 210 in association with the doctor ID, it is not necessary to store all of the information. (See FIG. 7).
- FIG. 11 is a flowchart showing a processing procedure of the pathological medical terminal 220 according to the present embodiment. This flowchart is executed by the CPU 910 in FIG. 9 using the RAM 940, and realizes the function of the pathologist terminal 220 in FIG.
- step S1101 the pathology slide is read by the scanner 222 at a resolution corresponding to a magnification that can be analyzed.
- step S1103 an image corresponding to the read pathological slide is displayed on the display 223.
- step S1105 the pathologist waits for an ROI setting input. If there is an ROI setting input, the process proceeds to step S1207 to select a tissue area to be analyzed from the tissue specimen image corresponding to the pathological slide, and the ROI from the tissue area. Is set, and the priority of ROI is set.
- FIG. 4 shows an example in which a plurality of ROIs selected for the analysis request are displayed superimposed on the selected tissue region as a result of the processing in step S1207.
- step S1109 the read tissue specimen image data, the set ROI, and the priority order of the ROI are transmitted to the analysis center 210 with a doctor ID and an image ID.
- the terminal ID for identifying the pathologist terminal 220 and information related to the analysis method of the analysis center 210 are also transmitted together.
- step S1111 the reception of the analysis result and the similarity from the analysis center 210 is awaited. If the analysis result and the similarity are received, the process proceeds to step S1113 to generate a display screen in which the analysis result and the similarity are superimposed on the analysis target tissue region.
- step S1115 the display screen generated by superimposing is displayed on the display 223, thereby supporting the diagnosis of the pathologist (see FIG. 5).
- the comparison target for calculating the similarity with the ROI transmitted from the pathological medical terminal 220 is the ROI set by the analysis center 210.
- the analysis center 210 transmits the received tissue specimen image to the specialist medical terminal, requests ROI setting, and calculates the similarity with the specialist's ROI setting.
- the pathologist can evaluate his / her proficiency level and can also interact with ROI setting criteria.
- FIG. 12 is a block diagram illustrating a configuration of an information processing system 1200 including an analysis center 1210 that is an information processing apparatus according to the present embodiment.
- An information processing system 1200 in FIG. 12 includes an analysis center 1210 that is an information processing apparatus that supports pathological image diagnosis, and a plurality of pathological medical terminals 220 that receive support for pathological image diagnosis connected to the analysis center 1210 via a network 230. And a specialized medical terminal 1220 that supports pathological image diagnosis.
- a communication control unit 1215 for communicating with the network 230 controls communication with the specialized medical terminal 1220 in addition to the plurality of pathological medical terminals 220.
- the ROI receiving unit 1213 receives the ROI information from the requesting professional medical terminal 1220 received by the communication control unit 1215 together with the ROI analysis priority.
- the priority order of the ROI set by the specialist is hereinafter referred to as the second priority order.
- the ROI feature quantity analysis unit 214 analyzes the ROI feature quantity indicated by the ROI reception unit 211 and the ROI reception unit 1213 among the tissue specimen images stored in the tissue specimen image storage unit 212.
- the similarity calculation unit 216 uses the ROI and first priority of the ROI receiving unit 211, the ROI and second priority of the ROI receiving unit 1213, and the feature amount of each ROI analyzed by the ROI feature amount analysis unit 214. Thus, the similarity (correlation) between the ROI of the ROI receiving unit 211 and the ROI of the ROI setting unit 213 is calculated.
- the similarity calculated by the similarity calculation unit 216 is transmitted to the diagnosis support request source by the similarity / ROI transmission unit 217 via the communication control unit 1215 together with the ROI of the ROI setting unit 213. Further, since the similarity calculated by the similarity calculation unit 216 is referred to for subsequent diagnosis support, it is stored in the similarity storage DB 1218 so as to be searchable including information on the specialist to be compared in the request source unit.
- the specialized medical terminal 1220 includes a control unit 1221 that controls the operation of the specialized medical terminal 1220 and the communication with the analysis center 210. Further, a display 1223 for displaying the tissue specimen image transmitted from the analysis center 210 is provided.
- FIG. 12 does not show a keyboard or pointing device for data input or operation instruction, but it is assumed that necessary input / output devices are connected.
- the subject of the ROI setting request is a specialist, but it may be a pathologist who is not another specialist or another analysis center.
- FIG. 13 is a sequence diagram showing an operation procedure 1300 of the information processing system having the information processing apparatus according to the present embodiment.
- FIG. 13 operations from reading a pathological slide by the scanner 222 of the pathological medical terminal 220 to displaying the diagnosis support information on the screen will be described. Steps similar to those in FIG. 3 are denoted by the same reference numerals, and description thereof is omitted.
- step S 305 the tissue specimen image, the selected ROI, and the ROI priority order transmitted from the pathological medical terminal 220 to the analysis center 1210 are received by the analysis center 1210.
- the analysis center 1210 requests the specialist medical terminal 1220 to set the ROI.
- the ROI setting request is attached with a tissue specimen image and an image ID for specifying the tissue specimen image.
- the professional medical terminal 1220 displays the tissue specimen image received from the analysis center 1210 in step S1303.
- the specialist's ROI setting and ROI priority order are set, and the setting result is reported to the analysis center 1210 in step S1207.
- the analysis center 1210 registers request information (such as a tissue specimen image and ROI) from the pathological medical terminal 220 in step S1309 and waits for a report from the specialized medical terminal 1220. If there is a report of the priority of ROI and ROI from the specialized medical terminal 1220, the feature amount of each ROI is analyzed in step S309.
- the processing up to step S317 is the same as that in FIG. 3 except that the comparison target is changed from the ROI set by the analysis center to the ROI set by the specialist.
- FIG. 14 is a diagram showing an analysis result display screen 1400 on the display 223 of the pathological medical terminal 220 according to the present embodiment. 14 are the same as those in FIG. 5 on the display screen 1400, and therefore the description thereof will be omitted.
- FIG. 14 in addition to the information 510 in FIG. The name of the subject specialist who calculated the similarity is added. Note that the information shown in 1410 is an example, and the present invention is not limited to this.
- the pathologist can recognize the evaluation of the setting of his / her ROIs 401 to 404 in FIG.
- FIG. 15 is a diagram showing a configuration of the similarity accumulation DB 1218 according to the present embodiment.
- the difference between the configuration of the similarity storage DB 1218 in FIG. 15 and FIG. 7 is only the addition of the specialist name 1209, and the same reference numerals as those in FIG. 7 are the same as those in FIG. To do.
- FIG. 16 is a flowchart illustrating a processing procedure of the information processing apparatus according to the present embodiment. This flowchart is executed by a CPU (not shown) using a RAM, and realizes each functional component of the analysis center 1210 of FIG.
- step S805 of FIG. 8 is replaced with steps S1601 to S1603. Accordingly, the processing of the other steps with the same reference number is the same, so the description thereof will be omitted, and only the differences will be described below.
- step S1601 After storing the information received from the pathologist terminal 220 in step S803, in step S1601, the tissue specimen image received from the pathologist terminal 220 is transmitted to the specialist medical terminal 1220 to request ROI setting by the specialist. In step S1603, it waits for reception of the report of the ROI setting from the specialized medical terminal 1220. If there is reception, the process proceeds to step S1605, and the priority order of the ROI and ROI received from the specialized medical terminal 1220 is stored. Then, the feature amount analysis and rank correlation coefficient (similarity) calculation of each ROI are executed in the same manner as in FIG. In step S813, the similarity and the specialist name calculated in the similarity storage DB 218 are recorded in association with the doctor ID.
- the correlation (similarity) between the ROI setting of the analysis center and the ROI setting of the specialist is calculated for a plurality of ROIs freely set on the tissue region selected by the pathologist. Calculated.
- the tissue region is divided into a plurality of blocks (corresponding to the ROI: see FIG. 7 and FIG. 8) in a grid-like predetermined region range in advance.
- the doctor selects an analysis target block as the ROI from the plurality of blocks. Accordingly, the calculation of the rank correlation coefficient (similarity) is also a calculation between two block selection ranks indicating which block is selected from a plurality of common blocks.
- the configuration and operation of the information processing system of this embodiment, and the configuration and operation of the pathological medical terminal, specialist medical terminal, and analysis center that constitute the information processing system are the same as those in the second and third embodiments.
- the difference between this embodiment and the second and third embodiments is that the ROI is set by selecting from a plurality of blocks obtained by dividing the tissue region into predetermined lattice-shaped regions in advance. Since different arithmetic expressions are used for the rank correlation coefficient (similarity) in FIG. 8, the differences will be described below.
- FIG. 17 is a diagram showing a display screen 1700 of the display 223 at the time when an ROI image is transmitted from the pathological medical terminal 220 according to the present embodiment.
- lattices 1711 having dimensions based on the magnification of the tissue specimen image and the like are previously superimposed.
- the priority order is displayed as a circle number as the ROI selected by the pathologist.
- Reference numeral 1710 denotes information including the request destination and the doctor ID of the request source.
- FIG. 18 is a diagram showing a display screen 1800 for displaying the analysis result and similarity on the display 223 of the pathological medical terminal 220 according to the present embodiment.
- a grid 1811 similar to that in FIG. 17 is superimposed on the display screen 1800.
- the priority order is displayed in circles as ROIs selected by the pathologist at 1801 to 1804 corresponding to 1701 to 1704 in FIG.
- ROI priorities selected by the analysis center or the specialist are displayed in 1803, 1802, 1805, and 1804 in dotted-line circle numbers.
- 1810 is information including the doctor ID of the report source or the request source and the calculated similarity.
- Block division and ROI setting of tissue specimen image are performed as follows.
- the tissue area selected from the tissue specimen image is divided into small areas (blocks).
- a typical method there is a method of setting a small region by lattice-like block division. Also in this case, there is a method in which an area where no tissue exists is excluded from the target in advance and the rest is set as the target area.
- block division there is no overlap of areas, but there may be overlap.
- the position information of the ROI is fixed and known in advance.
- the pathologist selects an important block to be analyzed and sets it as an ROI.
- the ROI selection criterion is used.
- the ROI selection criterion is created based on general pathological knowledge such as the density of cell nuclei and the shape of a set of cell nuclei, and artificial intelligence information created by a machine learning algorithm. Note that it is not necessary to rank all the blocks, and it is possible to rank only the upper fixed number of blocks and rank the least significant blocks for the remaining blocks. For example, in FIG. 17, all the blocks other than the priority orders 1 to 4 are set to the priority order 5.
- Spearman's rank correlation coefficient Rs is Defined by Here, N is the total number of classifications (all blocks: target area), and dn is the rank difference in the corresponding classification. Rs takes a numerical value from “ ⁇ 1” to “+1”. When it is close to “+1”, a positive correlation (correlation in the same order) is high, and when it is close to “ ⁇ 1”, a negative correlation (one side) Correlation in the reverse order) is high and close to “0” indicates no correlation.
- the rank difference dn includes the normalized distance between the blocks when the blocks are set.
- the rank matching includes the normalized distance between blocks at the time of setting the blocks. Since Spearman's rank correlation coefficient Rs and Kendall's rank correlation coefficient Rk have the same rank, the ROI selection list may have the same rank. Therefore, it is not necessary to rank all the blocks in the block division, and the rank number can be set to be equal to or less than the block number. For example, as described above, it is possible to rank only the upper fixed number of blocks and give the lowest rank to the remaining blocks. Furthermore, it is sufficient for the similar calculation only to obtain the rank correlation coefficient for the number of blocks obtained by adding the upper fixed number of selected blocks, and the calculation time can be greatly reduced.
- FIG. 19 is a block diagram illustrating a configuration of an information processing system 1900 having an analysis center 1910 that is an information processing apparatus according to the present embodiment.
- the configuration of FIG. 19 is different from the configuration of FIG. 12 in that the tendency evaluation unit 1919 that evaluates the ROI setting tendency of doctors and devices using the similarity stored in the similarity storage DB 1218 and the tendency evaluation unit 1919
- the communication control unit 1915 notifies the pathological medical terminal 220 and the specialized medical terminal 1220 of the evaluated tendency.
- the other functional components having the same reference numerals as those in FIG. 12 are the same as those in FIG.
- FIG. 20 is a flowchart illustrating a processing procedure of the analysis center 1910 that is the information processing apparatus according to the present embodiment. This flowchart is executed by the CPU using the RAM, and realizes each functional component of the analysis center 1910 of FIG.
- step S2001 a tissue specimen image, one ROI setting, and the priority order of the ROI are received.
- step S2003 another ROI setting based on the same tissue specimen image and the priority order of the ROI are received.
- the two settings may be doctors, doctors and devices, or devices.
- step S2005 a rank correlation coefficient is calculated. Any of [Equation 1] to [Equation 4] may be used for such calculation.
- [Formula 1] in the case of block division it may be considered that the center position of the ROI is determined in advance.
- step S2007 it is determined whether rank correlation coefficients have been calculated for all images to be evaluated.
- the entire image to be evaluated may be a plurality of tissue regions from one pathological slide, or may be different pathological slides of the same part in the case of education or learning. Furthermore, it is good also as an image input in the long term including time passage. If the rank correlation coefficient has not been calculated for all images to be evaluated, the process returns to step S2001 and the process is repeated.
- step S2009 the average value of the rank correlation coefficient is calculated.
- step S2011 the average value of the rank correlation coefficients is compared with a predetermined reference value. If the average value of the rank correlation coefficient is larger than the predetermined reference value, the process proceeds to step S2013, and both ROI settings are stored in the similarity storage DB 1218 in association with both comparison targets, assuming that both ROI settings have a similar tendency. If the average value of the rank correlation coefficients is equal to or smaller than the predetermined reference value, the process proceeds to step S2015, and both ROI settings have different tendencies and are stored in the similarity storage DB 1218 in association with both comparison targets.
- ROI selection can be used to determine whether or not to redo.
- FIG. 21 is a sequence diagram showing an operation procedure of the information processing system 2100 including the information processing apparatus according to the present embodiment.
- the configuration of the information processing system 2100 is the same as that of FIG. 19 of the fifth embodiment.
- step S2101. The pathologist terminal 220 requests the analysis center 1910 for a diagnosis tendency table.
- the analysis center 1910 creates a diagnosis tendency table with reference to the similarity accumulation DB 1218.
- step S2105 the generated diagnostic tendency table is returned to the pathological terminal 220.
- step S2107 the pathological medical terminal 220 displays the received diagnosis tendency table on the display 223, and selects a diagnosis support destination (a plurality of diagnosis support destinations are possible).
- the subsequent processing is the same as the diagnosis support request sequence to the specialist medical terminal 1220 in FIG. 13 of the third embodiment.
- the pathological medical terminal 220 reads a tissue specimen image from the pathological slide by the scanner 222.
- a tissue region to be used for diagnosis is selected from a plurality of tissue regions in the tissue specimen image.
- an ROI for requesting the analysis to the analysis center 210 for diagnosis support is selected from the selected tissue region.
- the pathological medical terminal 220 transmits the tissue specimen image, the selected ROI, and the diagnosis support destination to the analysis center 1910.
- step S2113 the tissue specimen image and the selected ROI transmitted from the pathological medical terminal 220 to the analysis center 1910 are received by the analysis center 1910.
- step S2115 the analysis center 1910 requests the specialist medical terminal 1220 corresponding to the diagnosis support destination to set the ROI.
- the ROI setting request is attached with a tissue specimen image and an image ID for specifying the tissue specimen image.
- the specialist medical terminal 1220 of the diagnosis support destination displays the tissue specimen image received from the analysis center 1910.
- step S2117 the ROI setting of the specialist and the priority order of the ROI are performed, and the setting result is analyzed in step S2119. Reported to Center 1910.
- the analysis center 1910 registers request information (such as a tissue specimen image and ROI) from the pathological medical terminal 220 in step S2121 and waits for a report from the diagnostic medical support professional medical terminal 1220. If there is a report of ROI and ROI priority from the specialized medical terminal 1220, the feature amount of each ROI is analyzed in step S2123. In step S2125, the analysis result is transmitted to the diagnosis support request source.
- request information such as a tissue specimen image and ROI
- the pathological medical terminal 220 that has received the analysis result in step S2125 superimposes the analysis result and the similarity on the tissue specimen image read from the pathological slide in step S2109, and displays the result on the display 223 in step S2129.
- the pathologist refers to the analysis result displayed on the display 223 as support information, diagnoses the tissue specimen image, and can evaluate and learn his / her ROI setting based on the similarity.
- FIG. 22 is a diagram showing a display screen 2200 for an inquiry response to the display 223 of the pathological medical terminal 220 according to the present embodiment.
- doctor 2205 having similar tendencies evaluated by the tendency evaluation unit 1919 are arranged in descending order of the average value of the rank correlation coefficient corresponding to the part 2203 of the living tissue. Further, doctors 2207 having different tendencies are arranged from the one with the lowest average value of the rank correlation coefficient.
- the display of the diagnosis tendency table 2201 enables the pathologist to select another pathologist who receives diagnosis support so that the diagnosis becomes more accurate.
- FIG. 23 is a flowchart illustrating a processing procedure of the analysis center 1910 that is the information processing apparatus according to the present embodiment. This flowchart is executed by the CPU using the RAM, and realizes each functional component of the analysis center 1910 of FIG.
- step S2301 it is determined whether there is a request for a diagnostic tendency table from the pathological medical terminal 220. If there is a request for a diagnostic tendency table from the pathological medical terminal 220, the process proceeds to step S2303, and the rank correlation coefficient stored in the similarity storage DB 1218 is searched based on the requested doctor ID. In step S2305, the average value of the retrieved rank correlation coefficients for each correlation target doctor is calculated. In step S2307, the calculated average value is compared with a predetermined reference value, and doctors with similar / different tendencies are determined as in steps S2011 to S2015 in FIG. 20, and a diagnostic tendency table is generated. . In step S2309, the generated diagnostic tendency table is transmitted to the requested pathological medical terminal 220.
- the present invention may be applied to a system composed of a plurality of devices, or may be applied to a single device. Furthermore, the present invention can also be applied to a case where a control program that realizes the functions of the embodiments is supplied directly or remotely to a system or apparatus. Therefore, in order to realize the functions of the present invention on a computer, a control program installed in the computer, a medium storing the control program, and a WWW (World Wide Web) server that downloads the control program are also included in the scope of the present invention. include.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Radiology & Medical Imaging (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
Description
生体組織を撮影した組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力手段と、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算手段と、
を備えることを特徴とする。
生体組織を撮影した組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力ステップと、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算ステップと、
を含むことを特徴とする。
生体組織を撮影した組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力ステップと、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算ステップと、
をコンピュータに実行させる制御プログラムを格納したことを特徴とする。
生体組織を撮影した組織標本画像から診断対象として選択し第1の複数領域を送信する送信手段と、
前記第1の複数領域の送信に応答して、前記第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域とから、各領域の組織標本画像の特徴量により重み付けされた、前記選択された各領域間の距離に基づいて演算された類似度と、前記第2の複数領域とを受信する受信手段と、
前記受信手段が受信した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示手段と、
を備えることを特徴とする。
生体組織を撮影した組織標本画像から診断対象として選択した第1の複数領域を送信する送信ステップと、
前記第1の複数領域の送信に応答して、前記第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域とから、各領域の組織標本画像の特徴量により重み付けされた、前記選択された各領域間の距離に基づいて演算された類似度と、前記第2の複数領域とを受信する受信ステップと、
前記受信ステップにより受信した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示ステップと、
を含むことを特徴とする。
生体組織を撮影した組織標本画像から診断対象として選択した第1の複数領域を送信する送信ステップと、
前記第1の複数領域の送信に応答して、前記第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域とから、各領域の組織標本画像の特徴量により重み付けされた、前記選択された各領域間の距離に基づいて演算された類似度と、前記第2の複数領域とを受信する受信ステップと、
前記受信ステップにより受信した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示ステップと、
をコンピュータに実行させる制御プログラムを格納したことを特徴とする。
生体組織を撮影した組織標本画像に基づく診断を支援する情報処理システムであって、
前記組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力手段と、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算手段と、
前記演算手段が演算した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示手段と、
を備えることを特徴とする。
生体組織を撮影した組織標本画像に基づく診断を支援する情報処理方法であって、
前記組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力ステップと、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算ステップと、
前記演算ステップにおいて演算した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示ステップと、
を含むことを特徴とする。
本発明の第1実施形態としての情報処理装置100について、図1を用いて説明する。図1に示すように、情報処理装置100は、入力部110と演算部120とを含む。入力部110は、生体組織を撮影した組織標本画像から選択された診断対象となる第1の複数領域と、組織標本画像から選択された診断対象となる第2の複数領域と、組織標本画像上での前記選択された各領域の位置情報とを入力する。演算部120は、第1の複数領域と第2の複数領域との類似度を、選択された各領域間の組織標本画像上の距離を考慮した相関に基づいて演算する。
本発明の第2実施形態は、組織標本画像の1つの組織領域内に自由に選択され、病理医用端末から送信された複数のROIについて、同じ組織領域について他のROIの選択結果との類似度(相関の強さ)を演算して、ROIの選択が旨くなされているかを評価する。第2実施形態では、他のROIの選択は解析センターの病理画像解析支援装置により自動的に行なわれる。従って、本実施形態によれば、病理医用端末から送信された複数のROIの選択を解析センターが定量的に評価できる。
図2は、本実施形態の情報処理装置である解析センター210を含む情報処理システム200の構成を示すブロック図である。図2の情報処理システム200は、病理画像診断の支援を行なう情報処理装置である解析センター210と、解析センター210にネットワーク230を介して接続する病理画像診断の支援を受ける複数の病理医用端末220とを備える。
図3は、本実施形態に係る情報処理装置を含む情報処理システムの動作手順300を示すシーケンス図である。図3においては、病理医用端末220のスキャナ222による病理スライドの読み取りから、診断支援情報の画面表示までの動作を説明する。
以下、本実施形態の処理におけるディスプレイ223の表示画面を、図4及び図5に従って説明する。
図4は、本実施形態に係る、病理医用端末からROI画像を送信する時点での、病理医用端末220のディスプレイ223に表示された画面400を示す図である。
図5は、本実施形態に係る、解析結果と類似度とを受信した時点での、病理医用端末220のディスプレイ223に表示された画面500を示す図である。
図6は、本実施形態に係る解析センター210のハードウェア構成を示すブロック図である。なお、図6には、1つの装置による構成を示したが、機能別の複数の装置により構成されてもよい。
図7は、本実施形態に係る類似度蓄積DB218の構成を示すブロック図である。
図8Aは、本実施形態に係る情報処理装置の処理手順を示すフローチャートである。このフローチャートは、図6のCPU610がRAM640を使用しながら実行して、図2の解析センター210の各機能構成部を実現する。
図8AのステップS807で解析される特徴量としては、以下のような特徴量がある。しかし、以下の例はHE染色法の場合の一例であってこれには限定されない。
f1) 核のサイズ、
f2) 大きい核の密度 = 大きい核の数 / 全核の数、
f3) 腺管に属する核の密度、
f4) 核の向き、
f5) 核の扁平度、
f6) 腺管の厚さ、
f7) 色(RGB)、
f8) 色(HSV)、
f9) 腺管領域、
f10) Gabor関数でフィルタリングした信号(方位特徴、配列)、
がある。
F1) 核のサイズ、
F2) 核の長径および短径、
F3) 円形度(円に近ければ最大値1を取り,円から外れている度合いが大きいほど小さい値)、
F4) テクスチャ、
F5) 色(RGB)、
F6) 色(HSV)、
F7) 腺管領域、
がある。
次に、図8AのステップS809において使用されるROI選択の相関値演算を説明する。
本実施形態では、病理画像における特徴に依存したROI選択の類似度を算出する順位相関係数Rlを導入する。この順位相関係数は、1からROIの重心間の距離数を引いたものに特徴量の関数である重み係数により重み付けしながら線形和を取ったものであり、下記の式で表される。
図9は、本実施形態に係る病理医用端末220のハードウェア構成を示すブロック図である。図2に示したように、病理医用端末220は、基本的構成として、制御部221と、スキャナ222と、ディスプレイ223とを有している。
図10は、本実施形態に係る画像識別テーブル942及び送受信データ943の構成を示す図である。
図11は、本実施形態に係る病理医用端末220の処理手順を示すフローチャートである。このフローチャートは、図9のCPU910がRAM940を使用しながら実行して、図2の病理医用端末220の機能を実現する。
第2実施形態においては、病理医用端末220から送信されたROIと類似度を算出する比較対象は解析センター210が設定したROIであった。本実施形態では、解析センター210ではなく、解析センター210は受信した組織標本画像を専門医用端末に送信してROI設定を依頼し、専門医のROI設定との類似度を算出する。本実施形態によれば、病理医側で自分の習熟度などが評価できると共に、ROI設定基準に関する対話なども可能となる。
図12は、本実施形態に係る情報処理装置である解析センター1210を有する情報処理システム1200の構成を示すブロック図である。
図13は、本実施形態に係る情報処理装置を有する情報処理システムの動作手順1300を示すシーケンス図である。図13においては、病理医用端末220のスキャナ222による病理スライドの読み取りから、診断支援情報の画面表示までの動作を説明する。なお、図3と同様のステップは同じ参照番号を付しており、説明は省略する。
以下、本実施形態の処理におけるディスプレイ223の表示画面を、図14に従って説明する。なお、病理医用端末220の診断支援の依頼時におけるディスプレイ223の表示は図4と同様であるので、説明を省略する。
図14は、本実施形態に係る病理医用端末220のディスプレイ223への解析結果の表示画面1400を示す図である。なお、図14の表示画面1400における図5と同じ参照番号を付された要素は同じものであるので、説明は省略する
図14においては、1410に、図5の510の情報に加えて、ROIの類似度を演算した対象の専門医の氏名が追加されている。なお、1410に示す情報は一例であって、これに限定されない。病理医はかかる表示を見て、図4の自分のROI401~404の設定についての評価を認識することができる。
図15は、本実施形態に係る類似度蓄積DB1218の構成を示す図である。なお、図15の類似度蓄積DB1218の構成の、図7との相違点は、専門医名1209の追加のみであり、他の図7と同じ参照番号は図7と同様であるので、説明は省略する。
図16は、本実施形態に係る情報処理装置の処理手順を示すフローチャートである。このフローチャートは、図示しないCPUがRAMを使用しながら実行して、図12の解析センター1210の各機能構成部を実現する。なお、図16のフローチャートにおいて、図8との相違点は、図8のステップS805をステップS1601~S1603に置き換えたものである。従って、他の同じ参照番号のステップの処理は同様であるので説明は省略し、以下相違点のみを説明する。
第2実施形態及び第3実施形態においては、病理医が選ばれた組織領域上に自由に設定した複数のROIについて、解析センターのROI設定や、専門医のROI設定との相関(類似度)を演算した。本実施形態においては、組織領域上に自由に設定した複数のROIではなく、組織領域を予め格子状の所定領域範囲の複数ブロック(ROIに相当:図7及び図8参照)に分割し、病理医はその複数ブロックから解析対象のブロックをROIとして選択する。従って、順位相関係数(類似度)の演算も共通の複数ブロックからどのブロックを選択したかの、2つのブロック選択順位間の演算となる。しかしながら、2つのブロック選択順位間の演算には、暗黙裏に格子状の所定領域範囲に分割された複数ブロックの組織領域上の規格化された位置関係が寄与していると考えられるので、病理画像の特徴の1つである位置関係が考慮されている。本実施形態によれば、第2及び第3実施形態よりも、より簡単な類似度の演算により病理医におけるROI設定を定量的に評価できる。
以下、本実施形態の処理におけるディスプレイ223の表示画面を、図17及び図18に従って説明する。
図17は、本実施形態に係る病理医用端末220からROI画像を送信する時点の、ディスプレイ223の表示画面1700を示す図である。
図18は、本実施形態に係る病理医用端末220のディスプレイ223への解析結果と類似度との表示画面1800を示す図である。
組織標本画像のブロック分割とROI設定は、以下のように行なわれる。
本実施形態で使用される類似度演算の順位相関係数の計算方法としては、スピアマンの順位相関係数と、ケンドールの順位相関係数とを説明する。
スピアマンの順位相関係数Rsは、
ケンドールの順位相関係数Rkは、
第2乃至第4実施形態においては、類似度を病理医に通知したり、類似度蓄積DBに累積したりする例を示した。本実施形態においては、類似度蓄積DBに累積した類似度から病理医や専門医、あるいは装置のROI設定の傾向を評価して、その評価結果を通知する。本実施形態によれば、医師や装置の傾向を知ることによって、教育や学習、あるいは複数の医師が診断に関与する場合の、医師の選択(傾向の異なる医師を選ぶなど)に応用が可能となる。
図19は、本実施形態に係る情報処理装置である解析センター1910を有する情報処理システム1900の構成を示すブロック図である。
図20は、本実施形態に係る情報処理装置である解析センター1910の処理手順を示すフローチャートである。このフローチャートは、CPUがRAMを使用しながら実行して、図19の解析センター1910の各機能構成部を実現する。
第5実施形態では、順位相関係数の演算毎に、両者の傾向(似た傾向か./異なる傾向か)を評価して通知した。本実施形態では、時間履歴を集積して、その傾向を診断傾向表として病理医用端末220から参照可能とした。本実施形態によれば、病理医が組織標本画像による診断支援を依頼する場合に、誰に対して支援を依頼するかを予め選択することができ、より効果的な支援が可能となる。
図21は、本実施形態に係る情報処理装置を含む情報処理システム2100の動作手順を示すシーケンス図である。情報処理システム2100の構成は、第5実施形態の図19と同様である。
図22は、本実施形態に係る病理医用端末220のディスプレイ223への問合せ回答の表示画面2200を示す図である。
図23は、本実施形態に係る情報処理装置である解析センター1910の処理手順を示すフローチャートである。このフローチャートは、CPUがRAMを使用しながら実行して、図19の解析センター1910の各機能構成部を実現する。
以上、実施形態を参照して本発明を説明したが、本発明は上記実施形態に限定されものではない。本発明の構成や詳細には、本発明のスコープ内で当業者が理解し得る様々な変更をすることができる。また、それぞれの実施形態に含まれる別々の特徴を如何様に組み合わせたシステム又は装置も、本発明の範疇に含まれる。
Claims (19)
- 生体組織を撮影した組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力手段と、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算手段と、
を備えることを特徴とする情報処理装置。 - 前記入力手段は、さらに前記第1の複数領域及び前記第2の複数領域の各々の領域から解析された特徴量を入力し、
前記演算手段は、さらに前記特徴量により重み付けされた相関に基づいて前記類似度を演算することを特徴とする請求項1に記載の情報処理装置。 - 前記入力手段に入力される前記第1の複数領域は前記組織標本画像から第1の優先順位を付けて選択され、前記第2の複数領域は前記組織標本画像から第2の優先順位を付けて選択されており、
前記演算手段は、前記第1の優先順位が付いた複数領域と前記第2の優先順位が付いた複数領域との類似度を演算することを特徴とする請求項1乃至4のいずれか1項に記載の情報処理装置。 - 前記入力手段に入力される前記第1の複数領域は前記組織標本画像を分割したブロックから第1の優先順位を付けて選択され、前記第2の複数領域は前記組織標本画像を分割した前記ブロックから第2の優先順位を付けて選択されており、
前記演算手段は、
Nを総分類(全ブロック:対象領域)数とし、
Fiを、系列1{Xi}の順位を昇順に並べ替え、対応する系列2{Yi}のYi(i=1,2,…,N-1/j=i+1,i+2,…,N)について、Yi<Yjなる個数とし、
Biを、Yi>Yjなる個数とし、
ΣFiは2変数の元の順序での方向が一致する回数とし、ΣBjを2変数の一方を逆順序にしたときの方向が一致する回数とする場合に、
ケンドールの順位相関係数、
- 前記第1の複数領域と、前記第2の複数領域とは、それぞれ病理医と病理画像解析装置とのいずれか選択し、
前記特徴量は、前記病理画像解析装置により前記第1の複数領域及び前記第2の複数領域の各々の領域から解析されることを特徴とする請求項1乃至7のいずれか1項に記載の情報処理装置。 - 前記病理画像解析装置は当該情報処理装置に含まれることを特徴とする請求項9に記載の情報処理装置。
- 前記演算手段が演算した類似度と該類似度の比較対象となった複数領域とを送信する第1送信手段をさらに備えることを特徴とする請求項1乃至9のいずれか1項に記載の情報処理装置。
- 前記演算手段が演算した類似度を前記入力手段への複数の入力元に対応付けて蓄積する蓄積手段と、
前記蓄積手段に蓄積した類似度の履歴に基づいて、前記複数の入力元における診断対象となる複数領域の選択を評価する評価手段と、
前記評価手段が評価した前記複数の入力元における選択の評価を送信する第2送信手段と、
をさらに備えることを特徴とする請求項1乃至10のいずれか1項に記載の情報処理装置。 - 前記評価手段は、前記蓄積手段に蓄積した類似度の履歴の平均値に基づいて、前記入力元における診断対象となる複数領域の選択の傾向を評価することを特徴とする請求項11に記載の情報処理装置。
- 生体組織を撮影した組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力ステップと、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算ステップと、
を含むことを特徴とする情報処理装置の制御方法。 - 生体組織を撮影した組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力ステップと、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算ステップと、
をコンピュータに実行させる制御プログラムを格納したことを特徴とする記憶媒体。 - 生体組織を撮影した組織標本画像から診断対象として選択し第1の複数領域を送信する送信手段と、
前記第1の複数領域の送信に応答して、前記第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域とから、各領域の組織標本画像の特徴量により重み付けされた、前記選択された各領域間の距離に基づいて演算された類似度と、前記第2の複数領域とを受信する受信手段と、
前記受信手段が受信した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示手段と、
を備えることを特徴とする情報処理装置。 - 生体組織を撮影した組織標本画像から診断対象として選択した第1の複数領域を送信する送信ステップと、
前記第1の複数領域の送信に応答して、前記第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域とから、各領域の組織標本画像の特徴量により重み付けされた、前記選択された各領域間の距離に基づいて演算された類似度と、前記第2の複数領域とを受信する受信ステップと、
前記受信ステップにより受信した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示ステップと、
を含むことを特徴とする情報処理装置の制御方法。 - 生体組織を撮影した組織標本画像から診断対象として選択した第1の複数領域を送信する送信ステップと、
前記第1の複数領域の送信に応答して、前記第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域とから、各領域の組織標本画像の特徴量により重み付けされた、前記選択された各領域間の距離に基づいて演算された類似度と、前記第2の複数領域とを受信する受信ステップと、
前記受信ステップにより受信した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示ステップと、
をコンピュータに実行させる制御プログラムを格納したことを特徴とする記憶媒体。 - 生体組織を撮影した組織標本画像に基づく診断を支援する情報処理システムであって、
前記組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力手段と、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算手段と、
前記演算手段が演算した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示手段と、
を備えることを特徴とする情報処理システム。 - 生体組織を撮影した組織標本画像に基づく診断を支援する情報処理方法であって、
前記組織標本画像から診断対象として選択された第1の複数領域と、前記組織標本画像から診断対象として選択された第2の複数領域と、前記組織標本画像上での前記選択された各領域の位置情報とを入力する入力ステップと、
前記第1の複数領域と前記第2の複数領域との類似度を、前記選択された各領域間の前記組織標本画像上の距離を考慮した相関に基づいて演算する演算ステップと、
前記演算ステップにおいて演算した類似度と、識別可能とした前記第1の複数領域及び前記第2の複数領域とを表示する表示ステップと、
を含むことを特徴とする情報処理方法
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/982,415 US9092851B2 (en) | 2011-01-31 | 2012-01-10 | Information processing system, information processing method, information processing apparatus, control method therefor, and storage medium storing control program therefor |
JP2012555770A JP5645146B2 (ja) | 2011-01-31 | 2012-01-10 | 情報処理システム、情報処理方法、情報処理装置及びその制御方法とその制御プログラム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2011-018936 | 2011-01-31 | ||
JP2011018936 | 2011-01-31 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2012105281A1 true WO2012105281A1 (ja) | 2012-08-09 |
Family
ID=46602502
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2012/050248 WO2012105281A1 (ja) | 2011-01-31 | 2012-01-10 | 情報処理システム、情報処理方法、情報処理装置及びその制御方法とその制御プログラムを格納した記憶媒体 |
Country Status (3)
Country | Link |
---|---|
US (1) | US9092851B2 (ja) |
JP (1) | JP5645146B2 (ja) |
WO (1) | WO2012105281A1 (ja) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2021506003A (ja) * | 2017-12-06 | 2021-02-18 | ベンタナ メディカル システムズ, インコーポレイテッド | デジタル病理学分析結果の格納および読み出し方法 |
JPWO2021117466A1 (ja) * | 2019-12-10 | 2021-06-17 | ||
JP6908806B1 (ja) * | 2020-09-16 | 2021-07-28 | BonBon株式会社 | プログラム、情報処理装置、方法 |
JP2022537542A (ja) * | 2019-06-18 | 2022-08-26 | ホアウェイ・テクノロジーズ・カンパニー・リミテッド | 動的な画像解像度評価 |
US11568657B2 (en) | 2017-12-06 | 2023-01-31 | Ventana Medical Systems, Inc. | Method of storing and retrieving digital pathology analysis results |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014080833A1 (ja) * | 2012-11-21 | 2014-05-30 | 株式会社東芝 | 超音波診断装置、画像処理装置及び画像処理方法 |
US20150124079A1 (en) * | 2013-11-06 | 2015-05-07 | Canon Kabushiki Kaisha | Image data forming apparatus and control method thereof |
AU2016211981A1 (en) * | 2015-01-31 | 2017-07-27 | F. Hoffmann-La Roche Ag | Quality control of automated whole-slide analysis |
JP6333871B2 (ja) * | 2016-02-25 | 2018-05-30 | ファナック株式会社 | 入力画像から検出した対象物を表示する画像処理装置 |
US10830102B2 (en) * | 2018-03-01 | 2020-11-10 | General Electric Company | Casing with tunable lattice structure |
CN110263791B (zh) * | 2019-05-31 | 2021-11-09 | 北京京东智能城市大数据研究院 | 一种识别功能区的方法和装置 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06251038A (ja) * | 1993-03-01 | 1994-09-09 | Toshiba Corp | 医用診断支援システム |
JPH07271872A (ja) * | 1994-03-30 | 1995-10-20 | Gakushiyuu Joho Tsushin Syst Kenkyusho:Kk | コンピュータによる画像診断学習支援方法 |
JP2000276587A (ja) * | 1999-01-22 | 2000-10-06 | Fuji Photo Film Co Ltd | 異常陰影検出処理方法およびシステム |
JP2007061607A (ja) * | 2005-08-30 | 2007-03-15 | Siemens Medical Solutions Usa Inc | 第1のオブジェクト及び1つ以上の他のオブジェクトを含む画像を処理するための方法及び画像データからの画像を処理するためのシステム |
JP2007286945A (ja) * | 2006-04-18 | 2007-11-01 | Fujifilm Corp | 類似画像検索装置および方法並びにプログラム |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0973116A1 (en) | 1993-03-01 | 2000-01-19 | Kabushiki Kaisha Toshiba | Medical information processing system for supporting diagnosis |
DE102004039191B4 (de) * | 2004-08-12 | 2007-09-27 | Siemens Ag | Verfahren und Vorrichtung zur Ermittlung und Überwachung von Parametern einer Bestrahlungstherapie |
US7580560B2 (en) * | 2005-07-18 | 2009-08-25 | Mitutoyo Corporation | System and method for fast template matching by adaptive template decomposition |
JP5245424B2 (ja) | 2008-01-25 | 2013-07-24 | 日本電気株式会社 | 病理組織画像撮影システム、病理組織画像撮影方法、および病理組織画像撮影プログラム |
JP5491836B2 (ja) * | 2009-01-30 | 2014-05-14 | 株式会社東芝 | 超音波診断装置、超音波画像処理装置、医用画像診断装置及び医用画像処理装置 |
US8831361B2 (en) * | 2012-03-09 | 2014-09-09 | Ancora Software Inc. | Method and system for commercial document image classification |
-
2012
- 2012-01-10 US US13/982,415 patent/US9092851B2/en active Active
- 2012-01-10 JP JP2012555770A patent/JP5645146B2/ja active Active
- 2012-01-10 WO PCT/JP2012/050248 patent/WO2012105281A1/ja active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06251038A (ja) * | 1993-03-01 | 1994-09-09 | Toshiba Corp | 医用診断支援システム |
JPH07271872A (ja) * | 1994-03-30 | 1995-10-20 | Gakushiyuu Joho Tsushin Syst Kenkyusho:Kk | コンピュータによる画像診断学習支援方法 |
JP2000276587A (ja) * | 1999-01-22 | 2000-10-06 | Fuji Photo Film Co Ltd | 異常陰影検出処理方法およびシステム |
JP2007061607A (ja) * | 2005-08-30 | 2007-03-15 | Siemens Medical Solutions Usa Inc | 第1のオブジェクト及び1つ以上の他のオブジェクトを含む画像を処理するための方法及び画像データからの画像を処理するためのシステム |
JP2007286945A (ja) * | 2006-04-18 | 2007-11-01 | Fujifilm Corp | 類似画像検索装置および方法並びにプログラム |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2021506003A (ja) * | 2017-12-06 | 2021-02-18 | ベンタナ メディカル システムズ, インコーポレイテッド | デジタル病理学分析結果の格納および読み出し方法 |
JP7197584B2 (ja) | 2017-12-06 | 2022-12-27 | ベンタナ メディカル システムズ, インコーポレイテッド | デジタル病理学分析結果の格納および読み出し方法 |
US11568657B2 (en) | 2017-12-06 | 2023-01-31 | Ventana Medical Systems, Inc. | Method of storing and retrieving digital pathology analysis results |
US11959848B2 (en) | 2017-12-06 | 2024-04-16 | Ventana Medical Systems, Inc. | Method of storing and retrieving digital pathology analysis results |
JP2022537542A (ja) * | 2019-06-18 | 2022-08-26 | ホアウェイ・テクノロジーズ・カンパニー・リミテッド | 動的な画像解像度評価 |
JP7188856B2 (ja) | 2019-06-18 | 2022-12-13 | ホアウェイ・テクノロジーズ・カンパニー・リミテッド | 動的な画像解像度評価 |
JPWO2021117466A1 (ja) * | 2019-12-10 | 2021-06-17 | ||
WO2021117466A1 (ja) * | 2019-12-10 | 2021-06-17 | Phcホールディングス株式会社 | 病理診断サポートシステムおよび病理診断サポート装置 |
JP7261319B2 (ja) | 2019-12-10 | 2023-04-19 | Phcホールディングス株式会社 | 病理診断サポートシステムおよび病理診断サポート装置 |
JP6908806B1 (ja) * | 2020-09-16 | 2021-07-28 | BonBon株式会社 | プログラム、情報処理装置、方法 |
JP2022049586A (ja) * | 2020-09-16 | 2022-03-29 | BonBon株式会社 | プログラム、情報処理装置、方法 |
Also Published As
Publication number | Publication date |
---|---|
JP5645146B2 (ja) | 2014-12-24 |
US9092851B2 (en) | 2015-07-28 |
US20130315456A1 (en) | 2013-11-28 |
JPWO2012105281A1 (ja) | 2014-07-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5645146B2 (ja) | 情報処理システム、情報処理方法、情報処理装置及びその制御方法とその制御プログラム | |
CN110148142B (zh) | 图像分割模型的训练方法、装置、设备和存储介质 | |
US20170053064A1 (en) | Personalized content-based patient retrieval system | |
US10496884B1 (en) | Transformation of textbook information | |
US11657487B2 (en) | Focus-weighted, machine learning disease classifier error prediction for microscope slide images | |
US10853449B1 (en) | Report formatting for automated or assisted analysis of medical imaging data and medical diagnosis | |
Huo et al. | AI applications in renal pathology | |
US8423596B2 (en) | Methods of multivariate data cluster separation and visualization | |
CN105612554B (zh) | 用于表征通过视频医疗设备获取的图像的方法 | |
US8711149B2 (en) | Graphical user interface for interpreting the results of image analysis | |
JP5716753B2 (ja) | 情報処理システム、情報処理方法、情報処理装置およびその制御方法とその制御プログラム | |
JP2018502279A (ja) | 組織学画像中の核の分類 | |
US20210090248A1 (en) | Cervical cancer diagnosis method and apparatus using artificial intelligence-based medical image analysis and software program therefor | |
CN112543934A (zh) | 一种确定异常程度的方法、相应的计算机可读介质和分布式癌症分析系统 | |
WO2017093524A1 (en) | Devices, systems and methods for automated quantitative scoring of digitized tissue images | |
Wang et al. | A deep learning framework design for automatic blastocyst evaluation with multifocal images | |
US20230147471A1 (en) | Systems and methods to process electronic images to determine salient information in digital pathology | |
Cesario et al. | Mobile support for diagnosis of communicable diseases in remote locations | |
Singh et al. | A novel hybridized feature selection strategy for the effective prediction of glaucoma in retinal fundus images | |
Olayiwola et al. | Lung-Related Diseases Classification Using Deep Convolutional Neural Network. | |
Schau et al. | Predicting primary site of secondary liver cancer with a neural estimator of metastatic origin | |
Lensink et al. | Segmentation of pulmonary opacification in chest ct scans of covid-19 patients | |
Shenouda et al. | Assessing robustness of a deep-learning model for COVID-19 classification on chest radiographs | |
Bhattacharyya et al. | Medical Image Analysis of Lung Cancer CT Scans Using Deep Learning with Swarm Optimization Techniques | |
Singh et al. | InceptionV3 in Medical Imaging: Enhancing Precision in Acute Lymphoblastic Leukaemia Diagnosis |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 12742256 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2012555770 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13982415 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 12742256 Country of ref document: EP Kind code of ref document: A1 |