WO2022259429A1 - 情報処理装置、情報処理方法、医療映像識別装置及びプログラムが格納された非一時的なコンピュータ可読媒体 - Google Patents
情報処理装置、情報処理方法、医療映像識別装置及びプログラムが格納された非一時的なコンピュータ可読媒体 Download PDFInfo
- Publication number
- WO2022259429A1 WO2022259429A1 PCT/JP2021/021954 JP2021021954W WO2022259429A1 WO 2022259429 A1 WO2022259429 A1 WO 2022259429A1 JP 2021021954 W JP2021021954 W JP 2021021954W WO 2022259429 A1 WO2022259429 A1 WO 2022259429A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- index
- elements
- series data
- information processing
- classes
- Prior art date
Links
- 230000010365 information processing Effects 0.000 title claims abstract description 78
- 238000003672 processing method Methods 0.000 title claims description 10
- 238000004364 calculation method Methods 0.000 claims abstract description 87
- 238000000034 method Methods 0.000 description 42
- 230000006870 function Effects 0.000 description 14
- 206010028980 Neoplasm Diseases 0.000 description 12
- 201000011510 cancer Diseases 0.000 description 12
- 230000015654 memory Effects 0.000 description 10
- 238000004891 communication Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 238000007689 inspection Methods 0.000 description 8
- 239000010750 BS 2869 Class C2 Substances 0.000 description 7
- 238000001514 detection method Methods 0.000 description 7
- 239000010749 BS 2869 Class C1 Substances 0.000 description 6
- 238000001617 sequential probability ratio test Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 238000002591 computed tomography Methods 0.000 description 3
- 238000002595 magnetic resonance imaging Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000002950 deficient Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 206010002329 Aneurysm Diseases 0.000 description 1
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/906—Clustering; Classification
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- 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
- G06T2207/30096—Tumor; Lesion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Definitions
- the present invention relates to an information processing device, an information processing method, a medical image identification device, and a non-transitory computer-readable medium storing a program.
- Patent Documents 1 to 4 disclose information processing techniques using a sequential probability ratio test (SPRT).
- SPRT is a type of technique for determining to which of a plurality of predetermined classes serial data that is sequentially input belongs.
- JP 2009-245314 A JP 2008-299589 A Japanese Patent Publication No. 2001-523824 International Publication No. 2020/194497
- the present disclosure has been made to solve such problems, and stores an information processing device, an information processing method, a medical image identification device, and a program that can accurately classify series data.
- the purpose is to provide a non-transitory computer-readable medium.
- the information processing apparatus includes acquisition means for sequentially acquiring a plurality of elements included in series data, and indicating to which of a plurality of classes each of the plurality of elements should belong.
- first calculating means for calculating an index based on two or more of the plurality of elements;
- second calculating means for calculating a weight indicating the importance of the index for each of the plurality of elements;
- a third calculation for calculating an integrated index indicating to which of the plurality of classes the series data should appropriately belong, by weighting the respective indices of the plurality of elements with the corresponding weights and integrating them.
- means, and classification means for classifying the series data into one of the classes based on the integrated index.
- An information processing method includes a step of sequentially acquiring a plurality of elements included in series data, and an index indicating to which of a plurality of classes each of the plurality of elements should belong. based on two or more of the plurality of elements; calculating a weight indicating the importance of the index for each of the plurality of elements; and a step of weighting the indexes with the corresponding weights and integrating them to calculate an integrated index indicating to which of the plurality of classes the series data should belong, based on the integrated index; and classifying the series data into one of the classes.
- the non-transitory computer-readable medium includes a step of sequentially obtaining a plurality of elements included in series data, and determining which of a plurality of classes each of the plurality of elements belongs to.
- an information processing device an information processing method, a medical image identification device, and a non-temporary computer-readable medium storing a program, which are capable of classifying series data with high accuracy.
- FIG. 1 is a schematic diagram showing the overall configuration of a series data classification system according to a first embodiment
- FIG. 3 is a functional block diagram of an information processing device provided in the series data classification system according to the first embodiment
- FIG. 6 is a flowchart showing an example of classification processing performed by the information processing apparatus according to the first embodiment
- FIG. 10 is a functional block diagram of a medical image identification device that is an application example of the information processing device according to the second embodiment
- FIG. 11 is a functional block diagram of an information processing device according to a third embodiment
- a sequence data classification system according to this embodiment will be described.
- the series data classification system of the present embodiment sequentially acquires and analyzes a plurality of elements included in the series data, thereby classifying the series data into one of a plurality of predetermined classes. System.
- series data means a data string that can be broken down into multiple elements.
- the series data may be time series data or non-time series data.
- time-series data include moving image data, audio data, and the like.
- non-time-series data include vegetation data sampled from multiple locations, inspection data of multiple locations on a product, and multiple biometric data for biometric authentication.
- the plurality of elements included in the series data may be a plurality of images (frames) forming the moving image.
- the multiple elements included in the series data may be inspection data of each part of the product. Note that series data and elements to which the classification processing of this embodiment can be applied are not limited to these.
- the classes classified by the series data classification system of the present embodiment are, for example, the first class indicating that the product is non-defective and the product is defective. It can be a second class indicating that When the series data are a plurality of images (frames) constituting medical data, the class classified by the series data classification system of the present embodiment is, for example, the first class indicating that the image contains a cancerous site. and a second class indicating that no cancerous site is included. Note that the number of classes may be three or more.
- FIG. 1 is a schematic diagram showing the overall configuration of the series data classification system according to this embodiment.
- FIG. 1 shows the configuration of hardware included in the series data classification system.
- the series data classification system includes an information processing device 100 , a data acquisition device 201 , an input device 202 and a display device 203 .
- the information processing device 100 is a computer such as a mobile phone, a smart phone, a desktop PC (Personal Computer), a laptop PC, or a server.
- the information processing apparatus 100 includes a processor 101 , a memory 102 , a storage 103 , an input/output I/F (Interface) 104 and a communication I/F 105 .
- Each unit of the information processing apparatus 100 is connected to each other via a bus, wiring, driving device, etc., and can mutually transmit and receive control signals and data.
- the processor 101 is, for example, an arithmetic processing device such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit).
- the memory 102 is, for example, a volatile or nonvolatile storage medium such as RAM (Random Access Memory) or ROM (Read Only Memory).
- the storage 103 is a nonvolatile storage medium such as an HDD (Hard Disk Drive), SSD (Solid State Drive), memory card, or the like.
- the memory 102 or storage 103 stores programs for realizing the information processing functions of the information processing apparatus 100 .
- the processor 101 may execute the program after reading it onto the memory 102 or may execute it without reading it onto the memory 102 .
- Non-transitory computer readable media includes various types of forms of storage media.
- Non-transitory computer-readable media include, for example, magnetic storage media, magneto-optical storage media, optical storage media, and semiconductor memory.
- Examples of magnetic storage media include flexible disks, magnetic tapes, and hard disk drives.
- An example of a magneto-optical storage medium is a magneto-optical disk.
- Examples of optical storage media include CD-ROM (Compact Disc Read Only Memory), CD-R (Compact Disc Recordable), and CD-R/W (Compact Disc Rewritable).
- Examples of semiconductor memory include mask ROM, PROM (Programmable ROM), EPROM (Erasable PROM), flash ROM, and RAM.
- the program may be supplied to the information processing apparatus 100 by various types of temporary computer-readable media.
- Transitory computer-readable media include, for example, electrical signals, optical signals, and electromagnetic waves.
- a temporary computer-readable medium can supply a program to the information processing apparatus 100 via a wired communication path such as an electric wire, an optical fiber, or a wireless communication path.
- the input/output I/F 104 is a communication interface for communicating with peripheral devices based on standards such as USB (Universal Serial Bus) and DVI (Digital Visual Interface).
- the input/output I/F 104 can communicate with the data acquisition device 201, the input device 202, and the display device 203 by wire or wirelessly. Accordingly, the information processing device 100 can transmit and receive data and control signals to and from the data acquisition device 201 , the input device 202 and the display device 203 .
- the communication I/F 105 is a communication interface based on standards such as Bluetooth (registered trademark), Wi-Fi (registered trademark), and 4G.
- the communication I/F 105 can establish a wired or wireless communication connection with an external device. Thereby, the information processing device 100 can transmit and receive data to and from an external device.
- the data acquisition device 201 is a device for acquiring series data.
- the data acquisition device 201 may be an inspection device provided in a factory or the like.
- the data acquisition device 201 may be a biometric information acquisition device such as a digital camera, a microphone, or a fingerprint scanner.
- the data acquisition device 201 is an endoscope system, an MRI (Magnetic Resonance Imaging) system, a CT (Computed Tomography) system, or other device for acquiring medical information.
- the data acquisition device 201 includes a device that acquires analog signals such as a sensor
- the data acquisition device 201 may include an AD conversion (Analog-to-Digital Conversion) device that converts analog signals into digital data.
- the series data acquired by the data acquisition device 201 is input to the information processing device 100 .
- the input device 202 is a user interface for accepting user's operation of the information processing device 100 .
- Examples of the input device 202 include a keyboard, mouse, trackball, touch sensor, pen tablet, and buttons.
- the display device 203 is a device that displays a screen based on drawing data processed by the processor 101 .
- Examples of the display device 203 include LCD (Liquid Crystal Display), CRT (Cathode Ray Tube) display, OLED (Organic Light Emitting Diode) display, and the like.
- the input device 202 and the display device 203 may be integrally formed as a touch panel.
- the hardware configuration shown in FIG. 1 is an example, and devices other than these may be added, and some devices may not be provided. Also, some devices may be replaced by other devices having similar functions. Furthermore, part of the functions of this embodiment may be provided by another device via a network, and the functions of this embodiment may be implemented by being distributed to a plurality of devices.
- the storage 103 may be replaced with cloud storage outside the information processing apparatus 100 .
- the data acquisition device 201 can be omitted if the acquisition of series data is performed by a system different from the series data classification system. Alternatively, the data acquisition device 201 , the input device 202 or the display device 203 may be provided within the information processing device 100 . Thus, the hardware configuration shown in FIG. 1 can be changed as appropriate.
- FIG. 2 is a functional block diagram of the information processing device 100 according to this embodiment.
- the information processing apparatus 100 includes an acquisition unit 110 , a first calculation unit 120 , a second calculation unit 130 , a third calculation unit 140 and a classification unit 150 .
- the first calculator 120 includes an index calculator 121 and a first storage 122 .
- the second calculator 130 includes a weight calculator 131 and a second storage 132 .
- the third calculator 140 includes an integrated index calculator 141 and a third storage 142 .
- the processor 101 implements the functions of the acquisition unit 110, the index calculation unit 121, the weight calculation unit 131, the integrated index calculation unit 141, and the classification unit 150 by executing programs stored in the memory 102, storage 103, or the like. Also, the processor 101 realizes the functions of the first storage unit 122, the second storage unit 132, and the third storage unit 142 by controlling the storage 103 based on the program. Specific processing performed by each of these units will be described later.
- FIG. 3 is a flowchart showing an example of classification processing performed by the information processing apparatus 100 according to this embodiment.
- the classification process shown in FIG. 3 is a process of classifying input series data into one of a plurality of predetermined classes.
- the classification process of FIG. 3 includes a loop process (steps S101 to S107) of acquiring elements one by one from series data including a plurality of elements and sequentially calculating integrated indices. This loop processing is repeated until the class into which the series data is to be classified is determined based on the integrated index.
- the processing in FIG. 3 may be started, for example, when a predetermined user operation is performed on the input device 202 .
- the processing start timing is not limited to this.
- the process of FIG. 3 may be executed when series data is input from the data acquisition device 201 .
- the processing in FIG. 3 may be repeated at predetermined time intervals. good.
- step S101 the acquisition unit 110 acquires one element of series data.
- the acquisition process at this time may directly acquire data from the data acquisition device 201, or may read data acquired in advance from the data acquisition device 201 and stored in the storage 103 or the like. After that, the process proceeds to step S102.
- the index calculation unit 121 refers to the first storage unit 122 and determines whether or not there is past data.
- the past data is a processing result (index, for example, likelihood ratio) by the index calculation unit 121 for an element obtained in the past, and one or more elements obtained in the past including the element.
- step S103 the index calculation unit 121 calculates an index indicating to which of a plurality of classes the input element should belong, considering the first element of the series data.
- the first calculation unit 120 outputs the calculated index to the second calculation unit 130, and stores the processing result in the first storage unit 122 as necessary.
- the index can be, for example, a likelihood ratio indicating the likelihood that a given element belongs to a given class among a plurality of classes.
- the index may be a function containing the likelihood ratio as a variable. In the following description, it is assumed that the index is the likelihood ratio. Note that the likelihood ratio calculated in step S103 is used as an integrated score. After that, the process proceeds to step S108. The processing of step S108 will be described later.
- step S104 the index calculation unit 121 reads past data stored in the first storage unit 122 .
- the first storage unit 122 stores a processing result and an input element each time the index calculation unit 121 performs processing. This storage process may overwrite previously stored information with new information, or may add new information while retaining previously stored information. After that, the process proceeds to step S105.
- the index calculation unit 121 indicates to which of a plurality of classes the input element should belong, considering two or more elements among the plurality of elements included in the series data. Calculate the index (likelihood ratio). Two or more elements include newly acquired elements and previously processed elements included in past data.
- the first calculation unit 120 outputs the calculated index to the second calculation unit 130, and stores the processing result in the first storage unit 122 as necessary.
- the index calculation unit 121 extracts features from the elements input from the series data. At this time, the index calculation unit 121 performs feature extraction in consideration of the input element when past data does not exist, and considers the relationship between the input element and the past data when past data exists. feature extraction.
- a convolutional neural network CNN
- LSTM Long Short Term Memory
- the probability that element x i (here, i is an arbitrary integer from 1 to N) belongs to class C 1 is p(x i
- these likelihood ratios are represented by the following formula (1).
- the likelihood ratio in Equation (1 ) indicates the likelihood ratio between the probability that element x i belongs to class C1 and the probability that element x i belongs to class C2 . For example, if the likelihood ratio exceeds 1, p(x i
- the index calculation unit 121 performs calculation taking into account the relationship between a plurality of elements, that is, the input elements and the past data as described above. be able to.
- the likelihood ratio calculated considering two elements x i and x j (i and j are arbitrary integers from 1 to N, where i ⁇ j) is given by the following equation (2): shall be written as
- step S106 the weight calculation unit 131 reads the likelihood ratio calculated in the past from the first storage unit 122 .
- the weight calculation unit 131 uses the likelihood ratio calculated this time and the likelihood ratio calculated in the past by the index calculation unit 121 to calculate the likelihood ratio calculated this time and the likelihood ratio calculated in the past.
- a weight corresponding to each of the degree ratios is calculated.
- the second storage unit 132 stores the weights calculated by the weight calculation unit 131 .
- the “weight” here is a value for adjusting the degree of influence of the likelihood ratio calculated this time and the likelihood ratio calculated in the past on the newly calculated integrated index. It is calculated according to the reliability of the selected element. “Reliability” is the degree of reliability related to the calculation of the likelihood ratio. degree becomes lower.
- the weight is, for example, a vector V ij whose elements are the probabilities that the input data belongs to each class, calculated considering the elements from the element x i to the element x j among the elements constituting the series data. and the inner product V ij V of the vector V iN whose element is the probability that the input data belongs to each class, calculated considering the elements from the element x i to the element x N among the elements constituting the series data iN .
- Both i and j are arbitrary integers from 1 to N. However, i ⁇ j.
- the weight is obtained by the inner product.
- the calculation of the weight is not limited to the inner product.
- an integrated index calculated in the past may be read from the third storage unit 142 and used to calculate the weight. After that, the process proceeds to step S107.
- step S107 the integrated index calculation unit 141 integrates the likelihood ratio calculated this time by the index calculation unit 121, the likelihood ratio calculated in the past, and the weight calculated this time by the weight calculation unit 131. Calculate a new integrated index.
- the integrated index indicates to which of multiple classes it is appropriate for the entire series data to belong.
- the past integrated index means the integrated index calculated by the integrated index calculation unit 141 for the element prior to the j-th element when this processing is for the j-th element of series data.
- the past likelihood ratio means the likelihood ratio calculated by the index calculation unit 121 for elements prior to the j-th element when this process is for the j-th element of series data.
- the third storage unit 142 stores the integrated index each time the integrated index calculation unit 141 performs processing. This memory processing may update the value of the integrated index by overwriting the previously stored integrated index with the new integrated index. An index may be added.
- the integrated index can be, for example, the sum of the values obtained by multiplying all the calculated likelihood ratios, including the past likelihood ratios, by the weight corresponding to each likelihood ratio.
- the integrated index may be a function including the integrated score as a variable. In the following description, it is assumed that the integrated index is the integrated score.
- the N elements are expressed as x 1 , . . . , x N .
- the probability that all the data including the elements from the i-th element to the N-th element of the sequence data belong to the class C 1 is expressed as p(x i , . . . , x N
- the probability that the entire data including the elements from the i-th element to the N-th element of the series data belongs to class C2 is expressed as p(x i , . . . , x N
- the integrated score is represented by the following formula (4) be.
- the integrated score is calculated by considering the importance of all the calculated likelihood ratios including the past likelihood ratios. Therefore, instead of using only one likelihood ratio as the integrated score as in Equation (4), the integrated score is calculated by different calculation formulas according to the number of all calculated likelihood ratios including past likelihood ratios. is calculated.
- the integrated score can be calculated using the following equation (5).
- i is an integer of 1 or more and N or less
- j is an integer of i or more and N or less.
- w ij is a weight calculated by the weight calculator 131 and indicating the importance of each likelihood ratio.
- i is an integer of 1 or more and N or less
- j is an integer of i or more and N or less.
- the probability that data containing only the i-th element belongs to class C 1 is p(x i
- the probability that data containing only the i-th element belongs to class C 2 Let be p(x i
- V ij be a vector whose elements are p(x i, . . . , x j
- w ij is the inner product V ij ⁇ V iN of V ij and V iN .
- the weight corresponding to the likelihood ratio when considering the i-th element to the j-th element of the sequence data considers the i-th element to the latest N-th element of the sequence data.
- the more similar the classification result is between the i-th element to the j-th element of the series data the larger the value.
- the likelihood ratio calculated taking into account the data inappropriate for discrimination is integrated as described above. The influence on the score can be reduced, and the discrimination accuracy can be improved.
- V N be a vector whose elements are probabilities p(x N
- C 2 ) are used as the weights w ij can also In this case, the most important weighting is the latest element x N of the series data.
- a process of normalizing the weights may be included.
- a normalization method for example, there is a method using the Softmax function.
- the likelihood ratio calculated in advance by the index calculation unit 121 in step S103 can be used as the likelihood ratio shown in Equation (5).
- Equation (5) shows an example of two-class classification in which the likelihood ratio between class C1 and class C2 is calculated, but the number of classes may be three or more.
- the formula (5) is extended so that the integrated score between the a-th class and all classes other than the a-th class among the M classes can be calculated. can be used.
- An example of such an extension is to use the maximum likelihood of all classes except the a-th class, as in Equation (6) below.
- the m-th class (m is an arbitrary integer from 1 to M) is C m
- a is an arbitrary integer from 1 to M
- b is an integer from 1 to M other than a.
- Equation (7) Another example is to use the sum of the likelihoods of all classes other than the a-th class, as in Equation (7) below.
- the m-th class (m is an arbitrary integer from 1 to M) is C m
- a is an arbitrary integer from 1 to M
- b is an integer from 1 to M other than a.
- Expressions (5) to (7) exemplify the case of using all of the input series data from the 1st element to the Nth element, but the number of elements to be considered may be arbitrary. For example, a maximum number P of elements to be used may be set in advance, and when the number of elements of input series data exceeds P, only the last P elements of the series data may be considered. This can prevent the calculation load from becoming too large.
- an integrated index may be calculated by a method using LSTM or a deep neural network.
- step S108 the classification unit 150 determines whether or not the series data can be classified into any class based on the integrated index calculated by the third calculation unit 140 .
- the classification unit 150 determines whether the class can be classified, for example, based on whether there is a class whose integrated score exceeds a predetermined threshold. If the classification is not possible (NO in step S108), the process proceeds to step S101, and the acquisition unit 110 acquires the next element. If the classification is possible (YES in step S108), the process proceeds to step S109.
- step S109 the classification unit 150 classifies the series data into one of the classes based on the integrated index. For example, if the integrated index is an integrated score, the series data is classified as belonging to a class whose integrated score exceeds a predetermined threshold.
- steps S108 and S109 will be described in more detail with a specific example.
- the classification process of this example is assumed to be a two -class classification into class C1 or class C2, and the threshold values used to determine class C1 and class C2 are T1 and T2, respectively .
- L be the integrated score.
- the classifying unit 150 classifies the series data into class C1, and the process ends. If L>T2, the classification unit 150 classifies the sequence data into class C2 , and the process ends. If L ⁇ T1 and L ⁇ T2, the classification unit 150 determines that classification is not possible, and the acquisition unit 110 acquires the next element.
- M thresholds are prepared in the same manner as described above, and the magnitude relationship between each of the M integrated scores and the corresponding threshold is determined.
- a similar classification process can be performed by At this time, the classification unit 150 classifies the series data into the class whose integrated score first exceeds the threshold. If the integrated score does not exceed any threshold, the classification unit 150 determines that classification is not possible, and the acquisition unit 110 acquires the next element.
- the above classification method is an example and is not limited to this. For example, if the number of elements input in steps S107 and S108 is greater than a predetermined value (maximum number of elements), even if there is no class whose integrated score exceeds the threshold, the series data is forced to either
- the procedure may be modified so as to end the processing after classifying the class. This can prevent the calculation time from becoming too long. In this example, it is desirable to make the criteria mutually exclusive to ensure classification into either class.
- series data is classified using a plurality of elements of series data and an integrated index that considers the importance of each element.
- classification can be performed according to the correlation length of the sequence data in consideration of the characteristics of the sequence data, so that the information processing apparatus 100 is provided that can classify the sequence data with high accuracy.
- a medical image identification device 300 will be described as one application example of the information processing device 100 of the first embodiment. Differences from the first embodiment will be mainly described below, and descriptions of common parts will be omitted or simplified.
- FIG. 4 is a functional block diagram of the medical image identification device 300 according to the second embodiment.
- the medical image identification device 300 includes a classification device 301 , a medical information acquisition unit 302 and a medical information storage unit 303 .
- the medical image identification device 300 can be configured including a computer, like the information processing device 100 shown in FIG. Therefore, description of the hardware configuration of the medical image identification device 300 is omitted.
- the medical image identification device 300 is, for example, a device that detects cancer from medical information such as endoscopic images, CT images, and MRI images.
- the medical image identification device 300 includes a device (such as an endoscope system) for acquiring medical information, and may operate standalone, and acquires medical information from other devices in the cancer detection system. cancer detection may be performed.
- the medical image identification device 300 may be composed of a plurality of devices that are communicatively connected to each other.
- the medical image identification device 300 can be, for example, an examination device for cancer examination. Alternatively, the medical image identification device 300 can be an inspection device for aneurysm inspection.
- the medical information acquisition unit 302 is a device that acquires medical information, and can be, for example, an endoscope system capable of capturing moving images.
- identifying medical information there is a case where a feature amount for matching is extracted from an image or the like acquired by the medical information acquisition unit 302 .
- This feature amount extraction processing may be performed in the classification device 301, may be performed in the medical information acquisition unit 302 when medical information is acquired, or may be performed by another device.
- the image itself acquired by the medical information acquisition unit 302 and the feature amount extracted therefrom may be collectively referred to as medical information.
- the medical information storage unit 303 stores information necessary for processing in the classification device 301 such as medical information.
- the information processing device 100 of the first embodiment is used for the classification device 301 .
- the classification device 301 acquires series data whose elements are medical information as the series data described in the first embodiment.
- the classification device 301 classifies the series data into one of a plurality of predetermined classes while referring to the information stored in the medical information storage unit 303 .
- the multiple classes may be, for example, classes indicating the presence or absence of cancer.
- the plurality of classes can include, for example, a class indicating that the input series data has a cancerous site and a class indicating that the input series data does not have a cancerous site.
- the medical image identification device 300 of this embodiment includes a classification device 301 capable of classifying series data with high accuracy. This provides the medical image identification device 300 that can detect cancer more appropriately.
- an example of cancer detection will be described as one example in which the feature of the information processing device 100 of the first embodiment, that is, the high accuracy of classification of series data is more utilized.
- machine learning is used to calculate the degree of certainty that cancer is included in each image (frame) included in the image, and the degree of certainty corresponding to one input image.
- a weighted sum of certainty factors corresponding to a predetermined number of fixed-length sheets is used as a classification score, and if the classification score exceeds a preset threshold, the image contains a cancerous site.
- a time-series image extracted from a medical video is input as series data, and the luminance value of the image is used as a feature value to perform class classification indicating the presence or absence of cancerous sites in the series data. Therefore, cancer detection is possible.
- a specific example of actually detecting cancer from an endoscopic image in the classification device 301 of this embodiment is shown below.
- the data at a certain point in the endoscopic image is regarded as the first element of the series data, and the elements up to N frames ahead are regarded as one series data. It should be classified as whether it is a non-cancerous site.
- N is an integer of 1 or more.
- step S101 one frame included in the endoscopic video is read from the medical information acquisition unit 302 or the medical information storage unit 303.
- step S102 it is checked whether there is a past element of the series data.
- step S103 calculate the likelihood ratio considering only the elements of the input series data, and set it as an integrated score.
- step S104 If there is a past element, proceed to step S104, read the past element, calculate the likelihood ratio considering the past data and the weight corresponding to the likelihood ratio in the procedure from step S104 to step S107, and calculate the integrated score Calculate
- step S108 the integrated score is compared with a threshold value for class discrimination, and if the integrated score exceeds the threshold value for discriminating cancerous sites, in step S109 the series data is treated as an image of cancerous sites. Classify. If the integrated score exceeds the threshold for discriminating non-cancerous sites, the series data is classified as non-cancerous site images in step S109. If the integrated score does not exceed the threshold value of any class, the process returns to step S101, and one frame included in the endoscopic video is newly acquired from the medical information acquisition unit 302 or the medical information storage unit 303. .
- the input sequence data includes cancerous sites that are easy to distinguish, cancerous sites that are difficult to distinguish, and non-cancerous sites.
- a cancerous site that can be easily identified is, for example, a cancerous site that protrudes like a lump.
- a difficult-to-distinguish cancerous site is, for example, a flat cancerous site with no protrusions or depressions. It is effective to detect cancerous sites that are difficult to distinguish using a plurality of images.
- the sequence data is long, there are cases where cancerous sites and non-cancerous sites are mixed in the sequence data.
- some of the elements of the series data that should be classified into different classes than the currently acquired elements are included. It is effective to set the weights so that the importance of the likelihood ratios calculated from the series data obtained is small, and to reduce the influence of the likelihood ratios on the integrated score for classification. Therefore, when classifying cancer detection, it is effective to use the classification processing of the information processing apparatus 100 of the first embodiment that uses one or more elements of series data while considering the importance of each element. is.
- the device or system described in the above embodiments can also be configured as in the following third embodiment.
- FIG. 5 is a functional block diagram of an information processing device 400 according to the third embodiment.
- the information processing apparatus 400 includes an acquisition unit 410 , a first calculation unit 420 , a second calculation unit 430 , a third calculation unit 440 and a classification unit 450 .
- Acquisition unit 410 sequentially acquires a plurality of elements included in series data.
- the first calculation unit 420 calculates, for each of the plurality of elements, an index indicating to which of the plurality of classes it is appropriate to belong, considering two or more of the plurality of elements.
- the second calculator 430 calculates a weight indicating the importance of each index of the plurality of elements.
- the third calculation unit 440 weights and integrates the respective indices of the plurality of elements with the corresponding weights, and calculates an integrated index indicating to which of the plurality of classes the series data should belong. do.
- the classification unit 450 classifies series data into one of classes based on the integrated index.
- an information processing device 400 that can classify series data with high accuracy.
- the present invention is not limited to the above-described embodiments, and can be modified as appropriate without departing from the gist of the present invention.
- an example in which a part of the configuration of one of the embodiments is added to another embodiment, or an example in which a part of the configuration of another embodiment is replaced is also an embodiment of the present invention.
- ASIC Application Specific Integrated Circuit
- FPGA Field Programmable Gate Array
- a floppy (registered trademark) disk, hard disk, optical disk, magneto-optical disk, CD (Compact Disk)-ROM, magnetic tape, nonvolatile memory card, and ROM can be used as the storage medium. Also, it is not limited to the one that executes the processing by the program recorded in the storage medium alone, but the one that operates on the OS (Operating System) and executes the processing in cooperation with other software and the functions of the expansion board. are also included in the scope of each embodiment.
- SaaS Software as a Service
- (Appendix 1) Acquisition means for sequentially acquiring a plurality of elements included in series data; a first calculation means for calculating, based on two or more of the plurality of elements, an index indicating which of the plurality of classes each of the plurality of elements should belong to; a second calculation means for calculating a weight indicating the importance of the index for each of the plurality of elements; a third step of calculating an integrated index indicating to which of the plurality of classes the series data should belong, by weighting the indexes of the plurality of elements with the corresponding weights and integrating them; calculating means; Classifying means for classifying the series data into one of the classes based on the integrated index; Information processing device.
- the second calculation means calculates a weight indicating the degree of importance of the index for each of the plurality of elements, one or more of the indicators including the index corresponding to the weight calculated by the first calculation means. calculated using The information processing device according to appendix 1.
- the second calculation means calculates a weight indicating the degree of importance of the index for each of the plurality of elements, the two or more indicators including the index corresponding to the weight calculated by the first calculation means. calculated using The information processing device according to appendix 1 or 2.
- the index includes a likelihood ratio that indicates the likelihood that each of the plurality of elements belongs to a certain class among the plurality of classes, 4.
- the information processing apparatus according to any one of Appendices 1 to 3.
- the integrated index includes an integrated score that indicates the likelihood that the series data belongs to a class among the plurality of classes, 5.
- the information processing apparatus according to any one of Appendices 1 to 4.
- the first calculation means is a first storage means for storing at least the elements acquired in the past by the acquisition means; When an element of the series data is newly acquired by the acquisition means, based on the newly acquired element and the previously acquired element stored in the first storage means, index calculation means for calculating the index for the newly acquired element; comprising The information processing apparatus according to any one of Appendices 1 to 8.
- the second calculation means is a second storage means for storing the index calculated in the past by the first calculation means; By using the index newly calculated by the first calculation means and the index calculated in the past stored in the second storage means, weight calculation means for calculating the weight; comprising The information processing device according to appendix 9.
- the third calculation means is The index newly calculated by the first calculation means, the previously calculated index stored in the second storage means, and the newly calculated index calculated by the weight calculation means and a weight for each of the indices calculated in the past; integrated index calculation means for calculating the integrated index; comprising 11.
- the series data is time series data, 12.
- the information processing apparatus according to any one of appendices 1 to 11.
- Appendix 13 medical information acquisition means for acquiring medical information of a subject;
- the information processing device according to any one of Appendices 1 to 12; with The information processing device classifies the series data containing the medical information as the element into one of the classes.
- Medical image identification device Medical image identification device.
- the information processing device classifies the series data into one of the classes indicating the presence or absence of cancerous sites in the medical information. 14.
- the medical image identification device according to appendix 13.
- (Appendix 15) a step of sequentially obtaining a plurality of elements included in the series data; calculating, based on two or more of the plurality of elements, an index indicating which of the plurality of classes each of the plurality of elements should belong to; calculating a weight indicating the importance of the index for each of the plurality of elements; a step of weighting the indexes of the plurality of elements with the corresponding weights and integrating them to calculate an integrated index indicating to which of the plurality of classes the series data appropriately belongs; , classifying the series data into one of the classes based on the integrated index;
- An information processing method comprising:
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- Primary Health Care (AREA)
- Epidemiology (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Radiology & Medical Imaging (AREA)
- Physics & Mathematics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Pathology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
Description
本実施形態に係る系列データ分類システムについて説明する。本実施形態の系列データ分類システムは、系列データに含まれる複数の要素を逐次的に取得して解析することにより、系列データをあらかじめ定められた複数のクラスのうちのいずれかに分類するためのシステムである。
本実施形態では、第1実施形態の情報処理装置100の適用例の1つとして、医療映像識別装置300を説明する。以下では主として第1実施形態との相違点について説明するものとし、共通部分については説明を省略又は簡略化する。
図5は、第3実施形態に係る情報処理装置400の機能ブロック図である。情報処理装置400は、取得部410、第1算出部420、第2算出部430、第3算出部440、及び分類部450を備える。取得部410は、系列データに含まれる複数の要素を逐次的に取得する。第1算出部420は、複数の要素の各々について、複数のクラスのいずれに属することが妥当であるかを示す指標を、複数の要素のうちの2以上の要素を考慮して算出する。第2算出部430は、複数の要素の各々の指標の重要度を示す重みを算出する。第3算出部440は、複数の要素のそれぞれの指標を対応する前記重みで重み付けしたうえで統合して、系列データが複数のクラスのいずれに属することが妥当であるかを示す統合指標を算出する。分類部450は、統合指標に基づいて、系列データをいずれかのクラスに分類する。
本発明は、上述の実施形態に限定されることなく、本発明の趣旨を逸脱しない範囲において適宜変更可能である。例えば、いずれかの実施形態の一部の構成を他の実施形態に追加した例や、他の実施形態の一部の構成と置換した例も、本発明の実施形態である。
系列データに含まれる複数の要素を逐次的に取得する取得手段と、
前記複数の要素の各々について、複数のクラスのいずれに属することが妥当であるかを示す指標を、前記複数の要素のうちの2以上の要素に基づいて算出する第1算出手段と、
前記複数の要素の各々の前記指標の重要度を示す重みを算出する第2算出手段と、
前記複数の要素のそれぞれの前記指標を対応する前記重みで重み付けしたうえで統合して、前記系列データが前記複数のクラスのいずれに属することが妥当であるかを示す統合指標を算出する第3算出手段と、
前記統合指標に基づいて、前記系列データをいずれかのクラスに分類する分類手段と、
を備える情報処理装置。
前記第2算出手段は、前記複数の要素の各々の前記指標の重要度を示す重みを、前記第1算出手段によって算出された、当該重みに対応する前記指標を含む1以上の前記指標、を用いて算出する、
付記1に記載の情報処理装置。
前記第2算出手段は、前記複数の要素の各々の前記指標の重要度を示す重みを、前記第1算出手段によって算出された、当該重みに対応する前記指標を含む2以上の前記指標、を用いて算出する、
付記1又は2に記載の情報処理装置。
前記指標は、前記複数の要素の各々が前記複数のクラスのうちのあるクラスに属することの尤もらしさを示す尤度比を含む、
付記1乃至3の何れか一項に記載の情報処理装置。
前記統合指標は、前記系列データが前記複数のクラスのうちのあるクラスに属することの尤もらしさを示す統合スコアを含む、
付記1乃至4の何れか一項に記載の情報処理装置。
前記分類手段は、前記統合スコアが所定の閾値を超えているクラスが存在する場合に、前記系列データを前記統合スコアが前記閾値を超えているクラスに分類する、
付記5に記載の情報処理装置。
前記統合スコアが所定の閾値を超えているクラスが存在しない場合に、前記分類手段は、前記系列データをいずれかのクラスにも分類せず、前記取得手段は、更に要素を取得する、
付記5又は6に記載の情報処理装置。
前記分類手段は、前記統合スコアが所定の閾値を超えているクラスが存在せず、かつ、前記取得手段によって取得された要素の数が所定値よりも多い場合に、前記統合スコアに基づいて前記系列データをいずれかのクラスに分類する、
付記5乃至7の何れか一項に記載の情報処理装置。
前記第1算出手段は、
前記取得手段によって過去に取得された前記要素を少なくとも記憶する第1記憶手段と、
前記取得手段によって前記系列データの要素が新たに取得されたときに、前記新たに取得された要素と、前記第1記憶手段に記憶されている過去に取得された前記要素と、に基づいて、前記新たに取得された要素に対する前記指標を算出する指標算出手段と、
を備える、
付記1乃至8の何れか一項に記載の情報処理装置。
前記第2算出手段は、
前記第1算出手段によって過去に算出された前記指標を記憶する第2記憶手段と、
前記第1算出手段によって新たに算出された前記指標と、前記第2記憶手段に記憶されている過去に算出された前記指標と、を使用することにより、使用された複数の前記指標の各々に対する前記重みを算出する重み算出手段と、
を備える、
付記9に記載の情報処理装置。
前記第3算出手段は、
前記第1算出手段によって新たに算出された前記指標と、前記第2記憶手段に記憶されている過去に算出された前記指標と、前記重み算出手段によって算出された、前記新たに算出された指標及び前記過去に算出された指標の各々に対する重みと、に基づいて、前記統合指標を算出する統合指標算出手段と、
を備える、
付記10に記載の情報処理装置。
前記系列データは、時系列データである、
付記1乃至11のいずれか1項に記載の情報処理装置。
対象者の医療情報を取得する医療情報取得手段と、
付記1乃至12のいずれか1項に記載の情報処理装置と、
を備え、
前記情報処理装置は、前記医療情報を前記要素として含む前記系列データをいずれかのクラスに分類する、
医療映像識別装置。
前記情報処理装置は、前記系列データを前記医療情報の癌化部位の有無を示すいずれかのクラスに分類する、
付記13に記載の医療映像識別装置。
系列データに含まれる複数の要素を逐次的に取得するステップと、
前記複数の要素の各々について、複数のクラスのいずれに属することが妥当であるかを示す指標を、前記複数の要素のうちの2以上の要素に基づいて算出するステップと、
前記複数の要素の各々の前記指標の重要度を示す重みを算出するステップと、
前記複数の要素のそれぞれの前記指標を対応する前記重みで重み付けしたうえで統合して、前記系列データが前記複数のクラスのいずれに属することが妥当であるかを示す統合指標を算出するステップと、
前記統合指標に基づいて、前記系列データをいずれかのクラスに分類するステップと、
を備える情報処理方法。
コンピュータに、
系列データに含まれる複数の要素を逐次的に取得するステップと、
前記複数の要素の各々について、複数のクラスのいずれに属することが妥当であるかを示す指標を、前記複数の要素のうちの2以上の要素に基づいて算出するステップと、
前記複数の要素の各々の前記指標の重要度を示す重みを算出するステップと、
前記複数の要素のそれぞれの前記指標を対応する前記重みで重み付けしたうえで統合して、前記系列データが前記複数のクラスのいずれに属することが妥当であるかを示す統合指標を算出するステップと、
前記統合指標に基づいて、前記系列データをいずれかのクラスに分類するステップと、
を備える情報処理方法を実行させるためのプログラムが格納された非一時的なコンピュータ可読媒体。
101 プロセッサ
102 メモリ
103 ストレージ
104 入出力I/F
105 通信I/F
110、410 取得部
120、420 第1算出部
121 指標算出部
122 第1記憶部
130、430 第2算出部
131 重み算出部
132 第2記憶部
140、440 第3算出部
141 統合指標算出部
142 第3記憶部
150、450 分類部
201 データ取得装置
202 入力装置
203 表示装置
300 医療映像識別装置
301 分類装置
302 医療情報取得部
303 医療情報記憶部
Claims (16)
- 系列データに含まれる複数の要素を逐次的に取得する取得手段と、
前記複数の要素の各々について、複数のクラスのいずれに属することが妥当であるかを示す指標を、前記複数の要素のうちの2以上の要素に基づいて算出する第1算出手段と、
前記複数の要素の各々の前記指標の重要度を示す重みを算出する第2算出手段と、
前記複数の要素のそれぞれの前記指標を対応する前記重みで重み付けしたうえで統合して、前記系列データが前記複数のクラスのいずれに属することが妥当であるかを示す統合指標を算出する第3算出手段と、
前記統合指標に基づいて、前記系列データをいずれかのクラスに分類する分類手段と、
を備える情報処理装置。 - 前記第2算出手段は、前記複数の要素の各々の前記指標の重要度を示す重みを、前記第1算出手段によって算出された、当該重みに対応する前記指標を含む1以上の前記指標、を用いて算出する、
請求項1に記載の情報処理装置。 - 前記第2算出手段は、前記複数の要素の各々の前記指標の重要度を示す重みを、前記第1算出手段によって算出された、当該重みに対応する前記指標を含む2以上の前記指標、を用いて算出する、
請求項1又は2に記載の情報処理装置。 - 前記指標は、前記複数の要素の各々が前記複数のクラスのうちのあるクラスに属することの尤もらしさを示す尤度比を含む、
請求項1乃至3の何れか一項に記載の情報処理装置。 - 前記統合指標は、前記系列データが前記複数のクラスのうちのあるクラスに属することの尤もらしさを示す統合スコアを含む、
請求項1乃至4の何れか一項に記載の情報処理装置。 - 前記分類手段は、前記統合スコアが所定の閾値を超えているクラスが存在する場合に、前記系列データを前記統合スコアが前記閾値を超えているクラスに分類する、
請求項5に記載の情報処理装置。 - 前記統合スコアが所定の閾値を超えているクラスが存在しない場合に、前記分類手段は、前記系列データをいずれかのクラスにも分類せず、前記取得手段は、更に要素を取得する、
請求項5又は6に記載の情報処理装置。 - 前記分類手段は、前記統合スコアが所定の閾値を超えているクラスが存在せず、かつ、前記取得手段によって取得された要素の数が所定値よりも多い場合に、前記統合スコアに基づいて前記系列データをいずれかのクラスに分類する、
請求項5乃至7の何れか一項に記載の情報処理装置。 - 前記第1算出手段は、
前記取得手段によって過去に取得された前記要素を少なくとも記憶する第1記憶手段と、
前記取得手段によって前記系列データの要素が新たに取得されたときに、前記新たに取得された要素と、前記第1記憶手段に記憶されている過去に取得された前記要素と、に基づいて、前記新たに取得された要素に対する前記指標を算出する指標算出手段と、
を備える、
請求項1乃至8の何れか一項に記載の情報処理装置。 - 前記第2算出手段は、
前記第1算出手段によって過去に算出された前記指標を記憶する第2記憶手段と、
前記第1算出手段によって新たに算出された前記指標と、前記第2記憶手段に記憶されている過去に算出された前記指標と、を使用することにより、使用された複数の前記指標の各々に対する前記重みを算出する重み算出手段と、
を備える、
請求項9に記載の情報処理装置。 - 前記第3算出手段は、
前記第1算出手段によって新たに算出された前記指標と、前記第2記憶手段に記憶されている過去に算出された前記指標と、前記重み算出手段によって算出された、前記新たに算出された指標及び前記過去に算出された指標の各々に対する重みと、に基づいて、前記統合指標を算出する統合指標算出手段と、
を備える、
請求項10に記載の情報処理装置。 - 前記系列データは、時系列データである、
請求項1乃至11のいずれか1項に記載の情報処理装置。 - 対象者の医療情報を取得する医療情報取得手段と、
請求項1乃至12のいずれか1項に記載の情報処理装置と、
を備え、
前記情報処理装置は、前記医療情報を前記要素として含む前記系列データをいずれかのクラスに分類する、
医療映像識別装置。 - 前記情報処理装置は、前記系列データを前記医療情報の癌化部位の有無を示すいずれかのクラスに分類する、
請求項13に記載の医療映像識別装置。 - 系列データに含まれる複数の要素を逐次的に取得するステップと、
前記複数の要素の各々について、複数のクラスのいずれに属することが妥当であるかを示す指標を、前記複数の要素のうちの2以上の要素に基づいて算出するステップと、
前記複数の要素の各々の前記指標の重要度を示す重みを算出するステップと、
前記複数の要素のそれぞれの前記指標を対応する前記重みで重み付けしたうえで統合して、前記系列データが前記複数のクラスのいずれに属することが妥当であるかを示す統合指標を算出するステップと、
前記統合指標に基づいて、前記系列データをいずれかのクラスに分類するステップと、
を備える情報処理方法。 - コンピュータに、
系列データに含まれる複数の要素を逐次的に取得するステップと、
前記複数の要素の各々について、複数のクラスのいずれに属することが妥当であるかを示す指標を、前記複数の要素のうちの2以上の要素に基づいて算出するステップと、
前記複数の要素の各々の前記指標の重要度を示す重みを算出するステップと、
前記複数の要素のそれぞれの前記指標を対応する前記重みで重み付けしたうえで統合して、前記系列データが前記複数のクラスのいずれに属することが妥当であるかを示す統合指標を算出するステップと、
前記統合指標に基づいて、前記系列データをいずれかのクラスに分類するステップと、
を備える情報処理方法を実行させるためのプログラムが格納された非一時的なコンピュータ可読媒体。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2021/021954 WO2022259429A1 (ja) | 2021-06-09 | 2021-06-09 | 情報処理装置、情報処理方法、医療映像識別装置及びプログラムが格納された非一時的なコンピュータ可読媒体 |
US18/567,102 US20240265536A1 (en) | 2021-06-09 | 2021-06-09 | Information processing apparatus, information processing method, medical image identification device, and non-transitory computer readable medium storing program |
JP2023526731A JPWO2022259429A5 (ja) | 2021-06-09 | 情報処理装置、情報処理方法、医療映像識別装置及びプログラム | |
EP21945105.1A EP4354322A4 (en) | 2021-06-09 | 2021-06-09 | INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING METHOD, MEDICAL IMAGE IDENTIFICATION DEVICE, AND NON-TRANSIENT COMPUTER-READABLE MEDIUM IN WHICH A PROGRAM IS STORED |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2021/021954 WO2022259429A1 (ja) | 2021-06-09 | 2021-06-09 | 情報処理装置、情報処理方法、医療映像識別装置及びプログラムが格納された非一時的なコンピュータ可読媒体 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022259429A1 true WO2022259429A1 (ja) | 2022-12-15 |
Family
ID=84425925
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/021954 WO2022259429A1 (ja) | 2021-06-09 | 2021-06-09 | 情報処理装置、情報処理方法、医療映像識別装置及びプログラムが格納された非一時的なコンピュータ可読媒体 |
Country Status (3)
Country | Link |
---|---|
US (1) | US20240265536A1 (ja) |
EP (1) | EP4354322A4 (ja) |
WO (1) | WO2022259429A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024154230A1 (ja) * | 2023-01-17 | 2024-07-25 | 日本電気株式会社 | 情報処理装置、情報処理方法、及び記録媒体 |
WO2024189833A1 (ja) * | 2023-03-15 | 2024-09-19 | 日本電気株式会社 | 情報処理装置、情報処理方法、及び情報処理プログラム |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001523824A (ja) | 1997-11-14 | 2001-11-27 | アーチ・デヴェロップメント・コーポレイション | スペクトル信号監視システム |
JP2008299589A (ja) | 2007-05-31 | 2008-12-11 | Hitachi Ltd | 生体認証システム |
JP2009245314A (ja) | 2008-03-31 | 2009-10-22 | Kddi Corp | 時系列データの識別装置および動画像への人物メタ情報付与装置 |
JP2010170280A (ja) * | 2009-01-21 | 2010-08-05 | Nippon Telegr & Teleph Corp <Ntt> | データ分類装置 |
WO2020194497A1 (ja) | 2019-03-26 | 2020-10-01 | 日本電気株式会社 | 情報処理装置、個人識別装置、情報処理方法及び記憶媒体 |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11527323B2 (en) * | 2019-05-14 | 2022-12-13 | Tempus Labs, Inc. | Systems and methods for multi-label cancer classification |
-
2021
- 2021-06-09 EP EP21945105.1A patent/EP4354322A4/en active Pending
- 2021-06-09 WO PCT/JP2021/021954 patent/WO2022259429A1/ja active Application Filing
- 2021-06-09 US US18/567,102 patent/US20240265536A1/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001523824A (ja) | 1997-11-14 | 2001-11-27 | アーチ・デヴェロップメント・コーポレイション | スペクトル信号監視システム |
JP2008299589A (ja) | 2007-05-31 | 2008-12-11 | Hitachi Ltd | 生体認証システム |
JP2009245314A (ja) | 2008-03-31 | 2009-10-22 | Kddi Corp | 時系列データの識別装置および動画像への人物メタ情報付与装置 |
JP2010170280A (ja) * | 2009-01-21 | 2010-08-05 | Nippon Telegr & Teleph Corp <Ntt> | データ分類装置 |
WO2020194497A1 (ja) | 2019-03-26 | 2020-10-01 | 日本電気株式会社 | 情報処理装置、個人識別装置、情報処理方法及び記憶媒体 |
Non-Patent Citations (1)
Title |
---|
See also references of EP4354322A4 |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024154230A1 (ja) * | 2023-01-17 | 2024-07-25 | 日本電気株式会社 | 情報処理装置、情報処理方法、及び記録媒体 |
WO2024189833A1 (ja) * | 2023-03-15 | 2024-09-19 | 日本電気株式会社 | 情報処理装置、情報処理方法、及び情報処理プログラム |
Also Published As
Publication number | Publication date |
---|---|
US20240265536A1 (en) | 2024-08-08 |
EP4354322A4 (en) | 2024-06-12 |
EP4354322A1 (en) | 2024-04-17 |
JPWO2022259429A1 (ja) | 2022-12-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP7248102B2 (ja) | 情報処理装置、個人識別装置、情報処理方法及び記憶媒体 | |
US11954852B2 (en) | Medical image classification method, model training method, computing device, and storage medium | |
US8290280B2 (en) | Image processing device, image processing method, and computer readable storage medium storing image processing program | |
WO2022259429A1 (ja) | 情報処理装置、情報処理方法、医療映像識別装置及びプログラムが格納された非一時的なコンピュータ可読媒体 | |
US9070041B2 (en) | Image processing apparatus and image processing method with calculation of variance for composited partial features | |
EP3998579A1 (en) | Medical image processing method, apparatus and device, medium and endoscope | |
WO2020232374A1 (en) | Automated anatomic and regional location of disease features in colonoscopy videos | |
US20220036140A1 (en) | Classification device, classification method, program, and information recording medium | |
US20140064563A1 (en) | Image processing apparatus, method of controlling image processing apparatus and storage medium | |
US20210407637A1 (en) | Method to display lesion readings result | |
US20140270499A1 (en) | Image processing apparatus, image processing method, and computer-readable recording device | |
US9201902B2 (en) | Techniques for medical image retrieval | |
CN110399907B (zh) | 基于诱导注意力的胸腔病症检测方法及装置、存储介质 | |
US9208173B1 (en) | Techniques for medical image retreival | |
US20220222820A1 (en) | Image processing apparatus, image processing method, and program | |
CN113827240B (zh) | 情绪分类方法和情绪分类模型的训练方法、装置及设备 | |
JP6425868B1 (ja) | 内視鏡画像観察支援システム、内視鏡画像観察支援装置、内視鏡画像観察支援方法 | |
US12087035B2 (en) | Information processing system, information processing method, and computer program | |
US20200167587A1 (en) | Detection apparatus and method and image processing apparatus and system, and storage medium | |
KR20230099995A (ko) | 자궁 경부암의 진단에 대한 정보 제공 방법 및 이를 이용한 자궁 경부암의 진단에 대한 정보 제공용 디바이스 | |
US20230401289A1 (en) | Information processing device, personal identification device, information processing method, and storage medium | |
US20230394117A1 (en) | Information processing device, personal identification device, information processing method, and storage medium | |
JP2022516139A (ja) | 残存がん細胞検出のための閾値化のためのシステムおよび方法 | |
JP2017152790A (ja) | 符号化装置、符号化方法、プログラム、及び画像処理システム | |
US20240119588A1 (en) | Image diagnosis support device, image diagnosis support method, remote diagnosis support system, and net contract service system |
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: 21945105 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2023526731 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2021945105 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2021945105 Country of ref document: EP Effective date: 20240109 |