Ultrasonic diagnostic device and model operation verification method

Abstract

To support detailed verification of an operation of an image analysis model.SOLUTION: A log managing unit 36 creates a log in which an operation of an ultrasonic diagnostic device is recorded. A model operation managing unit 40 preserves an image to be analyzed input to an image analysis model 30 and a result of the analysis of the image analysis model 30 for each operation of the image analysis model 30. When a specific model operation record in the log is selected, a report creation unit 42 creates a model operation report. The model operation report includes the image to be analyzed input to the image analysis model 30 and the result of the analysis of the image analysis model 30.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an ultrasound diagnostic apparatus and a model operation verification method, and in particular to a technique for verifying the operation of an image analysis model that is AI (Artificial Intelligence). [Background technology]

[0002] Ultrasound diagnostic devices are used in ultrasonic examinations of living bodies. Ultrasound diagnostic devices are medical devices that generate and display ultrasonic images based on reception signals obtained by transmitting and receiving ultrasonic waves. A typical ultrasonic image is a cross-sectional image.

[0003] Recently, ultrasonic diagnostic devices having an image analysis model generated through machine learning are becoming widespread. The ultrasonic diagnostic devices described in Patent Documents 1 and 2 have an image analysis model that automatically identifies (or classifies) tissue cross sections based on tomographic images. Other image analysis models include a model that detects a lesion included in a tomographic image, a model that measures tissue on a tomographic image, and the like.

[0004] Some ultrasound diagnostic devices have a log creation function. In such ultrasound diagnostic devices, a log is created by sequentially recording multiple operations in chronological order (see Patent Document 3). A series of past operations can be verified by referring to the log. Specifically, by referring to the log, it is possible to confirm that the operation of each component is appropriate, or to identify a malfunction of any of the components. In general, a log is composed of multiple records corresponding to multiple operations (including the occurrence of an event and the execution of a process) that occur in chronological order. The information contained in each record is finite and generally simple. Detailed information is not associated with each record in the log. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent No. 6423540 [Patent Document 2] JP 2020-204970 A [Patent Document 3] Special Publication No. 2021-507325 Summary of the Invention [Problem to be solved by the invention]

[0006] The performance of an image analysis model decreases due to various factors such as a decrease in the quality of the image to be analyzed and insufficient machine learning. It is desirable to periodically verify the performance or state of an image analysis model. Conventional ultrasound diagnostic devices do not fully meet such demands. For example, in an ultrasound diagnostic device having a log creation function, the operation of an image analysis model is recorded in a log. However, the contents of each record in the log are finite, and the contents of each record are generally simple. It is difficult to verify the operation of an image analysis model in detail by referring to each record.

[0007] An object of the present disclosure is to support verification of the operation of an image analysis model, or to support detailed verification of the operation of an image analysis model. [Means for solving the problem]

[0008] The ultrasound diagnostic device according to the present disclosure is characterized in that it includes an analysis unit having a machine-learned image analysis model that analyzes ultrasound images, a log creation unit that creates a log consisting of a plurality of action records by recording a plurality of actions including the actions of the image analysis model in chronological order, a storage unit that saves detailed information for model action verification for each action of the image analysis model, and a report creation unit that creates a model action report including detailed information associated with a specific model action record when the specific model action record in the log is selected.

[0009] The model operation verification method of the present invention is characterized by including the steps of: creating a log consisting of multiple operation records by recording in chronological order multiple operations, including operations of a machine-learned image analysis model that analyzes ultrasound images; saving detailed information for model operation verification for each operation of the image analysis model; displaying the log; accepting a selection of a specific model operation record in the displayed log; identifying detailed information associated with the specific model operation record; creating a model operation report including the detailed information; and displaying the model operation report. Effect of the Invention

[0010] According to the present disclosure, it is possible to assist in verifying the operation of an image analysis model. Alternatively, according to the present disclosure, it is possible to assist in detailed verification of the operation of an image analysis model. [Brief description of the drawings]

[0011] [Figure 1] 1 is a block diagram showing an example of the configuration of an ultrasound diagnostic apparatus according to an embodiment. [Diagram 2] 13 is a block diagram showing a configuration example of a model operation management unit. [Diagram 3] FIG. 13 is a diagram showing a detailed record column. [Figure 4] FIG. 13 illustrates an example of a log. [Diagram 5] FIG. 11 is a diagram showing a first example of a model operation report. [Figure 6] FIG. 13 is a diagram showing a second example of a model operation report. [Figure 7] FIG. 13 is a diagram showing a third example of a model operation report. [Figure 8] FIG. 13 is a diagram showing a fourth example of a model operation report. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, an embodiment will be described with reference to the drawings.

[0013] (1) Overview of the embodiment An ultrasound diagnostic apparatus according to an embodiment has an analysis unit, a log creation unit, a storage unit, and a report creation unit. The analysis unit has a machine-learned image analysis model that analyzes ultrasound images. The log creation unit creates a log consisting of a plurality of action records by recording a plurality of actions including the action of the image analysis model in chronological order. The storage unit saves detailed information for model action verification for each action of the image analysis model. When a specific model action record in the log is selected, the report creation unit creates a model action report including detailed information associated with the specific model action record.

[0014] According to the above configuration, the performance or state of the image analysis model can be verified in detail through reference to the detailed information included in the model operation report. The detailed information is more detailed than the information included in the log. For example, the model operation record in the log includes information such as the execution time of the image analysis model and the identifier of the image analysis model. When the amount of information is defined as a first amount of information, the detailed information has a second amount of information larger than the first amount of information. The detailed information according to the embodiment includes at least one of the analyzed image input to the image analysis model and the analysis result of the image analysis model. Usually, such information is discarded, but in the embodiment, such information is saved as detailed information for model operation verification. More specifically, the detailed information is saved while being associated with the model operation record in the log. Such an association is not recognized in conventional general logs.

[0015] In an embodiment, the storage unit stores the analyzed image as the above detailed information for each operation of the image analysis model. The analyzed image is an ultrasound image input to the image analysis model or an image equivalent thereto. The model operation report includes the analyzed image. With this configuration, it is possible to verify the operation of the image analysis model while referring to the analyzed image input to the image analysis model. All ultrasound images input to the image analysis model may be stored as the analyzed images, or a portion of all ultrasound images input to the image analysis model may be stored as the analyzed images.

[0016] The ultrasound diagnostic device according to the embodiment includes a conversion unit that converts an ultrasound image input to an image analysis model into a low-resolution image. The analyzed image is a low-resolution image. With this configuration, it is possible to reduce the storage capacity required to store the analyzed image. A storage period may be set for each analyzed image, and the analyzed image may be automatically deleted when the storage period has elapsed.

[0017] In an embodiment, the storage unit stores the analysis results of the image analysis model as the above detailed information for each operation of the image analysis model. The model operation report includes the analysis results. With this configuration, it is possible to verify the operation of the image analysis model while referring to the analysis results of the image analysis model. All the analysis results of the image analysis model may be stored, or a portion of all the analysis results of the image analysis model may be stored.

[0018] In an embodiment, the analysis result includes a plurality of scores corresponding to a plurality of classes. The report creation unit creates a graph based on the plurality of scores. The model operation report includes the graph. This configuration makes it easy to verify the operation of the image analysis model.

[0019] In an embodiment, the storage unit includes an image storage unit and a record storage unit. The image storage unit stores an analyzed image for each operation of the image analysis model. The record storage unit stores a detailed record including an analysis result of the image analysis model for each operation of the image analysis model. The analyzed image is a low-resolution image generated from an ultrasound image input to the image analysis model. The detailed information in the model operation report includes the analyzed image, which is input information for the image analysis model, and the analysis result, which is output information for the image analysis model.

[0020] According to the above configuration, the operation of the image analysis model can be verified while referring to both the input information and the output information of the image analysis model. Therefore, the performance and state of the image analysis model can be correctly identified, and if there is any problem, the cause can be identified. For example, if it is found that the image analysis model is not functioning properly, the image analysis model may be re-learned or replaced.

[0021] In an embodiment, each model operation record included in the log includes a portion in which a hyperlink is embedded to identify the location of detailed information associated with the record. The model operation report is displayed by operating the portion. With this configuration, when a model operation record of interest is found while referring to the log, it is possible to display the model operation report by performing a simple operation on the record. In this case, the model operation report may be displayed as a pop-up. In other words, the log and the model operation report may be displayed simultaneously.

[0022] The model operation verification method according to the embodiment includes a log creation process, a saving process, a log display process, a selection process, a specification process, a report creation process, and a display process. In the log creation process, a plurality of operations, including an operation of an image analysis model that has been machine-learned to analyze an ultrasound image, are recorded in chronological order, thereby creating a log consisting of a plurality of operation records. In the saving process, detailed information for model operation verification is saved for each operation of the image analysis model. In the log display process, the log is displayed. In the selection process, a selection of a specific model operation record in the displayed log is accepted. In the specification process, detailed information associated with the specific model operation record is specified. In the report creation process, a model operation report including the detailed information is created. In the display process, the model operation report is displayed.

[0023] A program for executing the above-mentioned model operation verification method is installed in an information processing device via a network or via a portable storage medium. The program is stored in a non-transitory storage medium in the information processing device. The information processing device is, for example, an ultrasound diagnostic device.

[0024] (2) Details of the embodiment An ultrasound diagnostic apparatus according to an embodiment is shown in Fig. 1. This ultrasound diagnostic apparatus is a medical apparatus that is installed in a medical institution such as a hospital and used for ultrasound examination of subjects.

[0025] The ultrasonic probe 10 is a device that transmits ultrasonic waves into a living body and receives reflected waves from the living body. The ultrasonic probe 10 has a transducer array consisting of a plurality of transducers. An ultrasonic beam 12 is formed by the transducer array. A beam scan plane 14 is formed by electronically scanning the ultrasonic beam 12. The beam scan plane 14 is repeatedly formed by repeating the electronic scanning of the ultrasonic beam 12 according to the transmission frame rate.

[0026] Known electronic scanning methods of the ultrasonic beam 12 include an electronic linear scanning method and an electronic sector scanning method. A two-dimensional transducer array may be provided as the transducer array. Volume data can be obtained from a three-dimensional space in a living body by two-dimensionally scanning the ultrasonic beam using the two-dimensional transducer array.

[0027] The transmission circuit 16 is an electronic circuit that functions as a transmission beamformer, and outputs a plurality of transmission signals in parallel to the transducer array during transmission, whereby a transmission beam is formed by the action of the transducer array.

[0028] The receiving circuit 18 is an electronic circuit that functions as a receiving beamformer, and applies phasing addition to multiple receiving signals output in parallel from the transducer array during reception, thereby generating receiving beam data. As electronic scanning is repeated, a receiving frame data string is output from the receiving circuit 18. Each receiving frame data is composed of multiple receiving beam data aligned in the electronic scanning direction. Each receiving beam data is composed of multiple echo data aligned in the depth direction.

[0029] The received frame data sequence is sent to the image forming unit 20 through a data processing unit (not shown). The data processing unit is a module that applies multiple processes to each received beam data. The multiple processes include logarithmic transformation, filtering, etc.

[0030] The image forming unit 20 has a digital scan converter (DSC). The DSC has a coordinate conversion function, a pixel interpolation function, etc. A display frame data string is generated from the received frame data string by the DSC. The display frame data string corresponds to a tomographic image as a moving image. Each display frame data constituting the display frame data string corresponds to a tomographic image as a still image.

[0031] In the illustrated configuration example, a display frame data sequence is temporarily stored in a cine memory 22, and then the display frame data sequence read out from there is sent to a display 26 via a display processing unit 24. A tomographic image is displayed as a moving image on the display 26, or a tomographic image is displayed as a still image. The display processing unit 24 has an image synthesis function, a color processing function, and the like.

[0032] The cine memory 22 is a temporary storage memory having a ring buffer structure. The cine memory 22 is composed of, for example, a semiconductor memory. The image forming unit 20 and the display processing unit 24 are each composed of, for example, a processor. The display 26 is composed of a liquid crystal display, an organic EL display device, or the like. The image forming unit 20 may form ultrasound images other than tomographic images. For example, blood flow images, elasticity images, and the like may be formed.

[0033] The image analysis unit 28 has an image analysis model 30 which is an AI. The image analysis model 30 is a machine-learned model, and is configured by, for example, a CNN (Convolutional Neural Network). The image analysis model 30 according to the embodiment is a model which identifies a tissue cross section based on a tomographic image. Specifically, it is a model which performs class classification. As the image analysis model 30, a model which performs measurement on tissue based on a tomographic image, a model which detects a lesion included in a tomographic image, or the like may be used.

[0034] In the embodiment, the display frame data string is transferred from the cine memory 22 to the image analysis unit 28. The display frame data string is also transferred to the information processing unit 34. The image analysis model 30 performs image analysis on each display frame data (i.e., each tomographic image) constituting the display frame data string, and outputs the analysis result for each display frame data. An analysis result string corresponding to the display frame data string is output from the image analysis unit 28 to the display processing unit 24 and the information processing unit 34. The analysis result string is sent to the display device 26 via the display processing unit 24, and the analysis result string is displayed on the display device 26. The image analysis unit 28 is composed of, for example, a processor. The received frame data may be analyzed in the image analysis unit 28.

[0035] The information processing unit 34 is composed of, for example, a CPU that executes a program. The information processing unit 34 functions as a control unit that controls the operation of each element that constitutes the ultrasound diagnostic device. In the figure, a plurality of functions performed by the information processing unit 34 are represented by a plurality of blocks. Specifically, the information processing unit 34 functions as a log management unit 36, a log analysis unit 38, and a model operation management unit 40. These will be described in detail later. An operation panel 50 is connected to the information processing unit 34. The operation panel 50 is an input device having a trackball, a plurality of switches, a plurality of knobs, etc.

[0036] A log storage unit 44, a detailed record storage unit 46, and an analyzed image storage unit 48 are connected to the information processing unit 34. These storage units 44, 46, 48 are configured with a semiconductor memory, a hard disk, or the like.

[0037] The log management unit 36 ​​functions as a log creation unit and a log reading unit. Specifically, a plurality of operations performed by the ultrasound diagnostic device are recorded in chronological order, and a log consisting of a plurality of records is created. Typically, one record corresponds to one operation. The plurality of operations performed by the ultrasound diagnostic device include a plurality of operations of an image analysis model. The plurality of records constituting the log include a plurality of records (a plurality of model operation records) corresponding to a plurality of model operations. In practice, the log management unit 36 ​​registers a plurality of records in chronological order in the log storage unit 44. As a result, a log is created on the log storage unit 44.

[0038] When referring to the log, the log management unit 36 ​​reads out a number of records registered in the log storage unit 44 and transfers these records to the display processing unit 24. As a result, these records are displayed as a log on the display. The user determines the display conditions. For example, the user specifies the display period. Alternatively, the display content is changed according to the user's scrolling operation.

[0039] The log analysis unit 38 analyzes the contents of the log. For example, the log analysis unit 38 has a counting function, a search function, and the like. The log analysis unit 38 may calculate the usage rate for each image analysis model, and may identify an operation related to the operation of the image analysis model (for example, input of approval or correction). The log analysis unit 38 may also calculate an inspection time, a processing speed, and the like. The log analysis unit 38 according to the embodiment has a report creation unit 42, which will be described later.

[0040] The model operation management unit 40 functions as a detailed information storage means. That is, the model operation management unit 40 stores the ultrasound image (specifically, a tomographic image) input to the image analysis model 30 as an analyzed image for each operation of the image analysis model 30, and also stores the analysis results (specifically, a plurality of scores) of the image analysis model 30. The analyzed image and the analysis results are basically information that is not included in the log, and are detailed information for verifying the model operation.

[0041] More specifically, the model operation management unit 40 generates a detailed record for each operation of the image analysis model 30. The detailed record includes link information indicating the location of the analyzed image and a numeric string as the analysis result.

[0042] A plurality of detailed records are stored in the detailed record storage unit 46. These records constitute a detailed record sequence. If the log stored in the log storage unit 44 is expressed as a main log or a general-purpose log, the detailed record sequence stored in the detailed record storage unit 46 can be called a sub-log or a log dedicated to an image analysis model.

[0043] The analyzed image storage unit 48 stores an analyzed image sequence consisting of a plurality of analyzed images. As described later, in an embodiment, each analyzed image is a low-resolution image generated from an ultrasound image. The detailed record storage unit 46 and the analyzed image storage unit 48 may be integrated. For example, the analyzed image may be included in each detailed record.

[0044] As described below, an ID number is issued for each action of an image analysis model to identify that action. Each model action record in the log contains an ID number, and each detailed record also contains an ID number. A specific detailed record is associated with each model action record through the ID number. The detailed record also contains link information that associates the detailed record with the image being analyzed.

[0045] When the user clicks on a specific model operation record in the displayed log, the report creation unit 42 obtains a specific detailed record associated with the specific model operation record, and also obtains a specific analyzed image based on the link information contained therein.

[0046] Then, the report creation unit 42 creates a model operation report based on the information included in the specific model operation record, the information included in the acquired detailed record, and the acquired analyzed image. The model operation report is displayed on the display 26. As will be described later, the model operation report includes the analyzed image as input information and the analysis results as output information. The user evaluates the performance and status of the image analysis model by referring to this information.

[0047] Fig. 2 shows a configuration example of the model operation management unit 40. In Fig. 2, the same components as those shown in Fig. 1 are denoted by the same reference numerals. The content shown in Fig. 2 corresponds to the algorithm of the model verification method according to the embodiment.

[0048] The information processing unit 34 includes a log management unit 36, a log analysis unit 38, and a model operation management unit 40. A log is constructed on a log storage unit 44. The model operation management unit 40 has a counter 52, a detailed record storage unit 54, a conversion unit 62, and an image storage unit 64. The detailed record storage unit 54 and the image storage unit 64 function as detailed information storage means 25. The image analysis unit 28 has an image analysis model 30. When a tomographic image 55 is input to the image analysis model 30, the tomographic image is analyzed by the image analysis model 30, and an analysis result 56 is output from the image analysis model 30.

[0049] The counter 52 issues an AI processing request identifier (hereinafter referred to as an ID number) for each operation of the image analysis model 30. When the image analysis model 30 performs a repeated operation, the counter 52 counts up, and the ID number increases sequentially. For example, the counter 52 may issue an ID number in response to a processing request from the main control unit (see reference numeral 58), or the counter 52 may issue an ID number in response to an input event of the tomographic image 55 (see reference numeral 60). The issued ID number is sent to the log management unit 36 ​​and the detailed record storage unit 54.

[0050] When creating a model action record, the log management unit 36 ​​includes an ID number in the model action record. When creating a detail record, the detail record storage unit 54 includes an ID number in the detail record. The ID number issued by the counter 52 may be sent to the log management unit 36 ​​and the detail record storage unit 54 via the image analysis unit 28. The analysis results of the image analysis model 30 are also sent to the detail record storage unit 54.

[0051] The conversion unit 62 applies low-resolution conversion to the tomographic image 55, thereby generating a low-resolution image. An example of the low-resolution image is a thumbnail image. The low-resolution conversion is a process for reducing the amount of information. The reduction rate can be 1 / 5, 1 / 10, or a rate lower than 1 / 10. Pixel thinning, compression, etc. may also be applied. From the viewpoint of verifying the image analysis model 30, both the tomographic image 55 and the low-resolution image can be referred to as the image to be analyzed. Hereinafter, the low-resolution image is referred to as the image to be analyzed.

[0052] The image saving unit 64 saves the analyzed image 66 in the analyzed image storage unit 48. The analyzed image storage unit 48 stores the analyzed images 66 in a number corresponding to the number of operations of the image analysis model 30. A storage period may be set for each analyzed image 66, and the analyzed image 66 may be deleted when the storage period has elapsed. Alternatively, the entire stored contents of the analyzed image 66 may be cleared according to an explicit instruction from the user. The image saving unit 64 notifies the detailed record saving unit 54 of information indicating the location of each analyzed image 66 as link information.

[0053] The detailed record saving unit 54 creates a detailed record including the ID number, link information, and analysis results, and stores it in the detailed record storage unit 46. The detailed record storage unit 46 stores a detailed record column 47. The detailed record column 47 is made up of a plurality of detailed records 57.

[0054] When a specific model action record in the log is selected by the user, the report creation unit 42 in the log analysis unit 38 retrieves a specific detailed record having the same ID number from the detailed record storage unit 46 based on the ID number contained in the specific model action record, and retrieves a specific analyzed image corresponding to the specific detailed record from the analyzed image storage unit 48 based on the link information contained in the specific detailed record. In reality, the model action record includes a portion in which a hyperlink that specifies the location of the detailed record is embedded. By clicking on that portion, the specific detailed record is immediately retrieved in the report creation unit 42.

[0055] As already described, the report creation unit 42 creates a model operation report including the information included in the specific model operation record, the information included in the specific detailed record, and the analyzed image. The model operation report is displayed on the display.

[0056] FIG. 3 shows the detailed record column 47. The detailed record column 47 is made up of a plurality of detailed records 57 corresponding to a plurality of operations of the image analysis model. Each detailed record 57 includes an ID number 67, link information (URL) 68, an image analysis model identifier 70, and an analysis result 72. The analysis result 72 is made up of a plurality of items 74 corresponding to a plurality of classes. Each item 74 is made up of a class identifier 76 and a score 78. The class identifier 76 indicates the type of tissue cross section to be identified. The score 78 is a confidence level. The content of the analysis result 72 may vary depending on the image analysis model used.

[0057] An example of a log is shown in Fig. 4. The log 80 reflects, for example, the process of executing a plurality of steps constituting an examination protocol. The log 80 is composed of a plurality of records (rows) arranged in chronological order. Basically, one record is generated for one operation (including an event, a state change, etc.) of the ultrasound diagnostic device. The log 80 includes a plurality of model operation records.

[0058] For example, model operation record 82 includes execution time 84, ID number 86, image analysis model identifier 88, representative analysis result 90, information 92 indicating whether the analysis was successful, etc. In part 94 corresponding to the ID number (part underlined and displayed in a specified hue), the URL of the detailed record is embedded as hyperlink information. By clicking on part 94, the corresponding detailed record is instantly identified, and a model operation report is created based on it, and the model operation report is displayed in a pop-up. For example, the model operation report is superimposed on the log.

[0059] Among the multiple records displayed, it is possible to easily identify each individual model operation record using the underline, ID number, etc. To enhance the identifiability or visibility of each individual model operation record, a special label may be added to each individual model operation record.

[0060] A first example of a model performance report is shown in Figure 5. The model performance report 96 includes text information 98, an analyzed image 100, and a graph 102. The text information 98 includes an execution time 104, an image analysis model identifier 106, an image analysis model version 108, a subject ID 110, a type of tissue cross section to be observed 124, and a representative analysis result 126.

[0061] The analyzed image 100 is a low-resolution image. In the illustrated example, a still image is displayed, but a moving image may also be displayed. The graph 102 is created based on a plurality of scores (a plurality of confidence levels) constituting the analysis result. The plurality of scores are arranged in ascending order from left to right, and the magnitude of each score is represented by the height of each bar. A plurality of classes (tissue cross-section types) corresponding to the plurality of scores are displayed along the horizontal axis (see reference numeral 128).

[0062] In the illustrated example, the type (class) of tissue cross section to be observed is S8 (see reference numeral 128a), and the corresponding score is relatively low (130a). This graph 102 indicates that acquisition of the desired tomographic image has failed (and is incomplete). Possible causes include the quality of the image to be analyzed, and also the performance of the image analysis model. If the content of the tomographic image is inappropriate due to an inappropriate position and orientation of the probe, the score of the class of interest will naturally be low. On the other hand, if the content of the tomographic image is appropriate and the score of the class of interest is low, it can be pointed out that the performance of the image analysis model may be insufficient.

[0063] In this way, by comparing the input information and output information contained in the model operation report, it is possible to comprehensively verify the operation of the image analysis model. The contents of the model operation report may be switched sequentially along the time axis. The model operation report is displayed as needed. In other words, if one simply wants to check the series of operations of the image analysis model, one can refer to the log itself.

[0064] A second example of the model operation report is shown in Fig. 6. For example, when an image analysis model for detecting a lesion is used, a model operation report 132 shown in Fig. 6 is displayed. A box 136 surrounding the detected lesion is shown in an analyzed image 134. This box 136 is a graphic element displayed together with a tomographic image. By referring to the display coordinates and size of the box 136, it is possible to verify whether the image analysis model has correctly detected the lesion. In this case, a moving image may be displayed instead of a still image.

[0065] FIG. 7 shows a third example of the model operation report. For example, when an image analysis model for measuring tissue is used, the model operation report 138 shown in FIG. 7 is displayed. The analyzed image 140 includes a tissue image 142 that is the measurement target. Two positions of interest in the tissue image 142 are automatically identified, and two markers 144, 146 are displayed for the two positions of interest. An auxiliary line 148 passing through them is also displayed. The distance between the two positions of interest is automatically measured on the auxiliary line 148. The result of the measurement is displayed as a numerical value (see reference numeral 150). For example, by comparing the positions of the tissue image 142 and the two markers 144, 146, it is possible to verify whether the operation of the image analysis model was appropriate.

[0066] Figure 8 shows a fourth example of a model operation report. The image analysis model is a model for classifying. Seven scores corresponding to seven tissue sections from S2 to S8 are calculated. Each score indicates the probability that the observed section corresponds to a specific tissue section.

[0067] The illustrated model operation report 152 includes multiple information sets 154-160 arranged in chronological order (see t1-t4). That is, this model operation report 152 corresponds to multiple model operation records, in other words, it is created based on multiple detailed records.

[0068] Each information set 154-160 is composed of an analyzed image 162 as input information and a graph 164 showing the analysis result as output information. Each graph 164 includes a plurality of bars expressing a plurality of scores. The bars filled with a dark color correspond to the tissue cross section (S8) to be observed. At t4, the analyzed image 162a is input to the image analysis model, and a good analysis result is obtained by the analysis. That is, in the graph 164a, a large value is calculated as the score corresponding to the tissue cross section (S8) to be observed. The height of the bar 168 showing that score is large, and the bar 168 is located at the top. By displaying such a model operation report, the temporal change in the operation of the image analysis model can be intuitively and easily understood.

[0069] As described above, according to the above embodiment, the performance or state of the image analysis model can be verified in detail by referring to the detailed information (input information and output information) included in the model operation report. Ultrasound images other than tomographic images may be analyzed. The operations of multiple image analysis models operating in parallel or selectively may be recorded. [Explanation of symbols]

[0070] 28 image analysis unit, 30 image analysis model, 36 log management unit, 38 log analysis unit, 40 model operation management unit, 42 report creation unit, 44 log storage unit, 46 detailed record storage unit, 48 analyzed image storage unit, 52 counter, 54 detailed record storage unit, 62 conversion unit, 64 image storage unit.

Claims

1. An analysis unit having a machine learning-processed image analysis model for analyzing ultrasound images; a log creation unit that creates a log including a plurality of action records by recording a plurality of actions including the action of the image analysis model in chronological order; A storage unit for storing detailed information for model operation verification for each operation of the image analysis model; a report creation unit that creates a model operation report including detailed information associated with a specific model operation record when the specific model operation record in the log is selected; 1. An ultrasonic diagnostic apparatus comprising:

2. 2. The ultrasonic diagnostic apparatus according to claim 1, The storage unit stores an analyzed image as the detailed information for each operation of the image analysis model, the image to be analyzed is an ultrasound image or an equivalent image input to the image analysis model; The model performance report includes the analyzed image.

1. An ultrasonic diagnostic apparatus comprising:

3. 3. The ultrasonic diagnostic apparatus according to claim 2, A conversion unit converts the ultrasound image input to the image analysis model into a low-resolution image, the analyzed image is the low resolution image; 1. An ultrasonic diagnostic apparatus comprising:

4. 2. The ultrasonic diagnostic apparatus according to claim 1, The storage unit stores an analysis result of the image analysis model as the detailed information for each operation of the image analysis model, The model operation report includes the analysis results.

1. An ultrasonic diagnostic apparatus comprising:

5. 5. The ultrasonic diagnostic apparatus according to claim 4, The analysis result includes a plurality of scores corresponding to a plurality of classes, the report creation unit creates a graph based on the plurality of scores; The model performance report includes the graph.

1. An ultrasonic diagnostic apparatus comprising:

6. 2. The ultrasonic diagnostic apparatus according to claim 1, The storage unit includes: an image storage unit that stores an analyzed image for each operation of the image analysis model; a record storage unit for storing a detailed record including an analysis result of the image analysis model for each operation of the image analysis model; Including, the analyzed image is a low-resolution image generated from an ultrasound image input to the image analysis model; The detailed information in the model operation report includes the analyzed image, which is input information of the image analysis model, and the analysis result, which is output information of the image analysis model.

1. An ultrasonic diagnostic apparatus comprising:

7. 2. The ultrasonic diagnostic apparatus according to claim 1, Each model action record included in the log includes an embedded hyperlink that identifies the location of the detailed information associated with the model action record; Selection of said portion displays said model performance report.

1. An ultrasonic diagnostic apparatus comprising:

8. A step of creating a log including a plurality of action records by recording a plurality of actions in chronological order, including an action of a machine-learned image analysis model that analyzes an ultrasound image; storing detailed information for model motion verification for each motion of the image analysis model; displaying the log; accepting a selection of a particular model action record in the displayed log; identifying detailed information associated with the particular model action record; generating a model performance report including said detailed information; displaying the model performance report; A model operation verification method comprising:

9. A program for executing a model operation verification method in an ultrasound diagnostic apparatus, comprising: The model operation verification method includes: A step of creating a log including a plurality of action records by recording a plurality of actions in chronological order, including an action of a machine-learned image analysis model that analyzes an ultrasound image; storing detailed information for model motion verification for each motion of the image analysis model; displaying the log; accepting a selection of a particular model action record in the displayed log; identifying detailed information associated with the particular model action record; generating a model performance report including said detailed information; displaying the model performance report; Including, A program characterized by:

Images