JP2024011742A - Medical image processing device, medical image processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the time and effort for re-segmentation of a medical image.

SOLUTION: A medical image processing device according to an embodiment comprises: an acquisition unit; a specification unit; a display control unit; a reception unit; and a re-segmentation unit. The acquisition unit acquires a medical image including an object region of a segmentation object and information about a segmentation result of the object region included in the medical image. The specification unit specifies an erroneous detection candidate region about the segmentation result. The display control unit displays information about the erroneous detection candidate region in association with the medical image. The reception unit receives labeling about the erroneous detection candidate region from a user. The re-segmentation unit executes segmentation of the object region on the basis of the labeling.

SELECTED DRAWING: Figure 2

COPYRIGHT: (C)2024,JPO&INPIT

Description

Embodiments disclosed in this specification and the drawings relate to a medical image processing device, a medical image processing method, and a program.

In order to segment medical images for use by users such as doctors, for example, there is a technique that applies a plurality of different segmentation methods to the medical image and allows the user to select a suitable segmentation method from the plurality of segmentation results. However, with this technique, if all or part of the segmentation results include many false positives, the user is required to perform re-segmentation.

Another technique is to automatically segment medical images, and if the model determines that the quality is low, it selects one representative slice with low quality, asks the user to input weakly supervised data for the representative slice, and responds to the input. There are techniques for correcting segmentation results. However, this technique only presents one representative slice with low quality. For this reason, when there are many false positives in the segmentation results, the user's effort to determine the position to be corrected increases.

Patent No. 6088498

The problem to be solved by the embodiments disclosed in this specification and the drawings is to reduce the effort required for resegmenting medical images. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problems. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems.

The medical image processing apparatus of the embodiment includes an acquisition section, a specification section, a display control section, a reception section, and a re-segmentation section. The acquisition unit acquires a medical image including a target region to be segmented, and information regarding a segmentation result of the target region included in the medical image. The specifying unit specifies a false detection candidate region regarding the segmentation result. The display control unit displays information regarding the erroneous detection candidate area in association with the medical image. The accepting unit accepts labeling regarding the false detection candidate area from the user. The re-segmentation unit performs segmentation of the target area based on the labeling.

FIG. 1 is a block diagram showing an example of the configuration of an in-hospital system 1 according to a first embodiment. FIG. 1 is a block diagram showing an example of the configuration of a medical image processing apparatus 100 according to the first embodiment. The figure which shows an example of the content of acquired image DB151. The figure which shows an example of the content of automatic segmentation result DB152. 5 is a flowchart illustrating an example of processing in the medical image processing apparatus 100. FIG. 3 is a diagram illustrating an overview of the process from grouping medical images to selecting a representative slice. FIG. 3 is a diagram showing an example of a visualized image displayed on the display 130. FIG. 3 is a diagram showing an example of a visualized image displayed on the display 130. The figure explaining the outline of the process of re-segmentation. FIG. 2 is a block diagram showing an example of the configuration of a medical image processing apparatus 100 according to a second embodiment. FIG. 3 is a diagram showing an example of a visualized image displayed on the display 130. A block diagram showing an example of the configuration of a medical image processing apparatus 100 according to the fifth embodiment

Hereinafter, a medical image processing apparatus, a medical image processing method, and a program according to embodiments will be described with reference to the drawings.

(First embodiment)
FIG. 1 is a block diagram showing an example of the configuration of an in-hospital system 1 according to the first embodiment. The in-hospital system 1 of the first embodiment includes, for example, a hospital information system (hereinafter referred to as HIS) 10, a radiology information system (hereinafter referred to as RIS) 20, and a medical image diagnostic apparatus (hereinafter referred to as RIS). modality) 30, a picture archiving and communication system (PACS) 40, and a medical image processing apparatus 100.

In the in-hospital system 1 , a medical image of a subject is captured by a modality 30 . A medical image captured by the modality 30 is automatically segmented in the modality 30. Automatic segmentation may be performed in a device other than modality 30, such as medical image processing device 100. In the medical image processing apparatus 100, re-segmentation is further performed based on the segmentation results performed by the modality 30 or other devices, and the final segmentation results are provided to the user together with the medical image. The provided medical images and segmentation results are used for interpretation and diagnosis by the user.

The HIS 10 is a computer system that provides business support within the hospital. Specifically, the HIS 10 has various subsystems. Various subsystems include, for example, an electronic medical record system, a medical accounting system, a medical reservation system, a hospital visit reception system, and an admission/discharge management system.

The HIS 10 is, for example, a computer such as a server device or a client terminal that is equipped with a processor such as a CPU (Central Processing Unit), a memory such as a ROM (Read Only Memory) or a RAM (Random Access Memory), a display, an input interface, and a communication interface. include.

A user uses the electronic medical record system included in the HIS 10 to input and refer to information regarding a patient. A user issues an image inspection order to the HIS 10. The HIS 10 transfers order information corresponding to the image inspection order to other systems such as the RIS 20.

RIS20 is a computer system that provides operational support in the image diagnosis department. The RIS 20 performs reservation management of image test orders in cooperation with the HIS 10, as well as coordination of reservation information with examination equipment and management of examination information. The RIS 20 includes, for example, a computer such as a server device or a client terminal that includes a processor such as a CPU, a memory such as a ROM or a RAM, a display, an input interface, and a communication interface.

The modality 30 executes imaging (photography) according to imaging conditions (imaging protocol) determined based on, for example, an image examination instruction. Examples of the modality 30 include an X-ray computed tomography device, an X-ray diagnostic device, a magnetic resonance imaging device, an ultrasound diagnostic device, a nuclear medicine diagnostic device, and the like. The modality 30 is operated by, for example, an operator such as a doctor (radiologist) or a medical radiology technician. A medical image (image data) generated by imaging by the modality 30 is transmitted to the PACS 40 and the medical image processing apparatus 100.

PACS 40 is a computer system that receives medical images transmitted by modality 30 and stores them in a database. The PACS 40 transmits (transfers) medical images stored in the database in response to requests from clients. The PACS 40 includes a server computer including a processor such as a CPU, a memory such as ROM and RAM, a display, an input interface, and a communication interface. The medical images stored in the PACS 40 are accompanied by information regarding the patient to be photographed and the photographing as additional information. The supplementary information includes information such as a patient ID, an examination ID, and imaging conditions (imaging protocol) in a format that complies with the DICOM (Digital Imaging and Communication in Medicine) standard, for example.

The configuration of the in-hospital system 1 is not limited to the above. The in-hospital system 1 may include, for example, an image interpretation report creation device. Moreover, some elements of the in-hospital system 1 may be integrated. For example, the HIS 10 and the RIS 20 may be integrated into one system.

FIG. 2 is a block diagram showing an example of the configuration of the medical image processing apparatus 100 of the first embodiment. Medical image processing apparatus 100 includes, for example, a communication interface 110, an input interface 120, a display 130, a processing circuit 140, and a memory 150.

The communication interface 110 communicates with external devices such as the RIS 20, the modality 30, and the PACS 40 via a network NW such as a LAN (Local Area Network). The communication interface 110 includes, for example, a communication interface such as a NIC (Network Interface Card). The network NW may include the Internet, a cellular network, a Wi-Fi network, a WAN (Wide Area Network), etc. instead of or in addition to a LAN.

The input interface 120 receives various input operations from a treating doctor and the like, converts the received input operations into electrical signals, and outputs the electrical signals to the processing circuit 140. For example, when an input operation is performed by a doctor or the like, the input interface 120 generates information according to the input operation.

Input operations include, for example, operations related to generation of order information, selection of reference image candidates, attachment of edited images, and editing of edited images. When the input operations in these cases are executed, the input interface 120 generates generation instruction information, attachment request information, selection information, and image editing information, respectively. The input interface 120 outputs information corresponding to the generated input operation to the processing circuit 140.

The input interface 120 includes, for example, a mouse, a keyboard, a trackball, a switch, a button, a joystick, a touch panel, and the like. The input interface 120 may be, for example, a user interface that accepts voice input such as a microphone. When the input interface 120 is a touch panel, the input interface 120 may also have the display function of the display 130.

Note that in this specification, the input interface is not limited to one that includes physical operation parts such as a mouse and a keyboard. For example, examples of the input interface include an electrical signal processing circuit that receives an electrical signal corresponding to an input operation from an external input device provided separately from the device and outputs this electrical signal to a control circuit.

Display 130 displays various information. For example, the display 130 displays images generated by the processing circuit 140, a GUI (Graphical User Interface) for accepting various input operations from an operator, and the like. For example, the display 130 is an LCD (Liquid Crystal Display), a CRT (Cathode Ray Tube) display, an organic EL (Electro Luminescence) display, or the like.

The processing circuit 140 includes, for example, an acquisition function 141, a specific function 142, a display control function 143, a reception function 144, and a re-segmentation function 145. The processing circuit 140 realizes these functions by, for example, a hardware processor (computer) executing a program stored in a memory (storage circuit) 150.

A hardware processor is, for example, a CPU, a GPU (Graphics Processing Unit), an Application Specific Integrated Circuit (ASIC), a programmable logic device (for example, a Simple Programmable Logic Device (SPLD), or Refers to circuits such as complex programmable logic devices (CPLDs) and field programmable gate arrays (FPGAs).

Instead of storing the program in the memory 150, the program may be directly incorporated into the circuit of the hardware processor. In this case, the hardware processor realizes its functions by reading and executing a program built into the circuit. The above program may be stored in advance in the memory 150 or in a non-temporary storage medium such as a DVD or CD-ROM, and the non-temporary storage medium may be a drive device ( (not shown) may be installed in the memory 150 from a non-transitory storage medium.

The hardware processor is not limited to being configured as a single circuit, but may be configured as one hardware processor by combining a plurality of independent circuits to realize each function. Further, a plurality of components may be integrated into one hardware processor to realize each function. Although the hardware processor, memory, etc. in the medical image processing apparatus 100 are provided separately from the hardware processor, memory, etc. in the HIS 10, they may be common.

The memory 150 stores, for example, an acquired image database (hereinafter referred to as DB) 151, an automatic segmentation result DB 152, and a segmentation model 153. The acquired image DB 151 is a DB of medical images acquired by the acquisition function 141. The automatic segmentation result DB 152 is a database of segmentation results transmitted by the modality 30, for example. The segmentation model 153 is a learned model generated by machine learning using segmentation results and supervised data with weakly supervised data (described later) as input data and re-segmentation results as output data.

FIG. 3 is a diagram showing an example of the contents of the acquired image DB 151. The acquired image DB 151 includes a plurality of medical images (slices). The medical image includes, for example, each item of patient ID, imaging date, modality, and image ID. The acquired image DB 151 includes a plurality of medical images for each patient ID, imaging date, and modality. The image ID item is assigned to each of a plurality of medical images, but indicates a common part of the image ID.

FIG. 4 is a diagram showing an example of the contents of the automatic segmentation result DB 152. The automatic segmentation result DB 152 includes a plurality of automatic segmentation results. The automatic segmentation result includes, for example, the following items: patient ID, image ID, image type, segmentation ID, and segmentation content.

The acquisition function 141 acquires information transmitted by external devices such as the HIS 10 , the RIS 20 , the modality 30 , and the PACS 40 and received by the communication interface 110 . For example, a plurality of medical images transmitted by the modality 30 are acquired. The medical image includes a target region for segmentation.

The target area is specified in advance in the medical image, for example, based on the electronic medical record transmitted by the HIS 10, for example, before performing automatic segmentation. The target area is arbitrary, and may be, for example, a body organ or a blood vessel. Medical images may be transmitted by a device other than modality 30, for example PACS 40. The acquisition function 141 stores the acquired medical images in the acquired image DB 151.

The acquisition function 141 acquires information regarding the segmentation results of the target area included in the medical image. The acquisition function 141 stores, for example, the automatic segmentation result transmitted by the modality 30 in the automatic segmentation result DB 152. The acquisition function 141 reads and acquires automatic segmentation results from the automatic segmentation result DB 152 stored in the memory 150. The acquisition function 141 is an example of an acquisition unit.

The identification function 142 identifies false positive candidate regions regarding the segmentation results. The identification function 142 identifies false detection candidate areas in a plurality of medical images acquired in one examination. Prior to specifying a false detection candidate area, the identification function 142 groups a plurality of medical images based on a classification criterion to create a false detection candidate group. The distribution criteria are determined using, for example, distances and image features within the medical image. Image characteristics include, for example, brightness and color.

The specific function 142 selects one or more representative slices from among the medical images included in the false detection candidate group. The specific function 142 adjusts the number of representative slices according to the number of medical images. For example, the specific function 142 adjusts the number of representative images to increase as the number of medical images increases. The specific function 142, for example, predetermines the number of medical images included in the false detection candidate group, adjusts the medical images to the determined number, and groups the plurality of medical images. The specific function 142 is an example of a specific part.

The display control function 143 causes the display 130 to display an image through display control. The display control function 143 causes the display 130 to display information regarding the erroneous detection candidate area specified by the specific function 142, in this case, the erroneous detection candidate area, in association with the medical image. In displaying the erroneous detection candidate area in association with a medical image such as a representative image, the display control function 143 causes the erroneous detection candidate area to be displayed in a superimposed manner (overlapping) with the medical image.

The display control function 143 may display the false detection candidate region in association with the medical image in another manner. For example, the display control function 143 may display the false detection candidate region and the medical image in association with each other by arranging them vertically or horizontally in an area having approximately the same area. Although the displayed medical image is a representative slice, it may be any other medical image. The display control function 143 causes the segmentation result to be displayed on the display 130 together with the false detection candidate area. The display control function 143 is an example of a display control section.

The reception function 144 receives labeling regarding the false detection candidate area from the user. Labeling includes, for example, weakly supervised data. Weakly supervised data is data that includes information regarding correction of segmentation results in false detection candidate regions. The user specifies correction of the segmentation result in the false detection candidate region by performing an input operation on the input interface 120.

The input interface 120 generates weakly supervised data in response to input operations on the input interface 120, and transmits the generated weakly supervised data to the processing circuit 140. The reception function 144 in the processing circuit 140 receives the weakly supervised data transmitted by the input interface 120. The reception function 144 is an example of a reception unit.

The re-segmentation function 145 performs segmentation of the target area in the medical image using the segmentation model 153 based on the segmentation results included in the automatic segmentation result DB 152 stored in the memory 150 and the weakly supervised data received by the reception function 144. (resegmentation).

The re-segmentation function 145 may perform re-segmentation using, for example, a graph cut, a region growing method, or the like, or may perform re-segmentation using a neural network. The re-segmentation function 145 is an example of a re-segmentation unit. Since the target region is a region set after automatic segmentation of a medical image, segmentation of the target region is re-segmentation.

Next, processing in the medical image processing apparatus 100 will be explained. FIG. 5 is a flowchart illustrating an example of processing in the medical image processing apparatus 100. The medical image processing apparatus 100 first determines whether a medical image to be subjected to automatic segmentation has been acquired by the acquisition function 141 (step S101). If it is determined that no medical image has been acquired, the acquisition function 141 repeats the process of step S101.

If it is determined that a medical image has been acquired, the acquisition function 141 stores the acquired medical image in the acquired image DB 151. Subsequently, the acquisition function 141 acquires the automatic segmentation result transmitted by the modality 30 (step S103). The target of segmentation may be specified by the user in advance, or may be specified after segmentation has been performed, for example. The acquisition function 141 stores the acquired segmentation results in the automatic segmentation result DB 152.

Subsequently, the identification function 142 identifies false detection candidate regions for each segmented medical image (step S105). For example, the specific function 142 calculates the predicted probability of the segmentation result for each pixel, identifies a false detection candidate region according to the height of the calculated predicted probability, and, for example, identifies a portion of pixels where the predicted probability is less than or equal to a predetermined percentage. Identify the area as a false detection candidate area. The identification function 142 may identify the false detection candidate area based on similar past segmentation results or image processing.

Subsequently, the specific function 142 groups the medical images based on the classification criteria determined using the distance, image characteristics, etc. in the medical images acquired by the acquisition function 141 (step S107), and creates false detection candidate groups. Create. The specific function 142 is defined as one group of medical images including false detection candidate areas that are clustered together between pixels and slices.

Subsequently, the specific function 142 selects a representative slice from among the plurality of medical images included in the created false detection candidate group (step S109). The specific function 142 may select the representative slice in any way. For example, the specific function 142 selects one representative slice when the number of medical images included in the false detection candidate group is a certain number or more, and selects a plurality of representative slices when the number exceeds a certain number.

When selecting one representative slice, the specific function 142 may select the medical image with the largest area of the segmentation result as the representative slice, or may select the central medical image when multiple medical images are arranged in chronological order. It may also be used as a representative slice. For example, when selecting a plurality of representative slices, the specific function 142 determines the number of representative slices according to the number of medical images included in the false detection candidate group, and for example, if the number of medical images included in the false detection candidate group is large. You can use as much as you like. For example, when selecting a plurality of representative slices, the specific function 142 may be a slice that divides the number of representative slices evenly when a plurality of medical images are arranged in chronological order.

FIG. 6 is a diagram illustrating an overview of the process from grouping medical images to selecting a representative slice. The upper part of the left column in FIG. 6 shows a plurality of medical images arranged in chronological order, and the lower part of the left column shows a plurality of medical images arranged in chronological order on which segmentation results are superimposed. False detection candidate regions are identified based on the plurality of medical images and segmentation results.

In the example shown in FIG. 6, three false detection candidate areas are identified, shown in the upper middle row, the middle row of the middle row, and the lower middle row, respectively. In the upper middle row, a false detection candidate area can be seen at the top center of the image. In the middle row and middle row, a false detection candidate area can be seen at the top left of the image. In the lower part of the middle row, a false detection candidate area can be seen in the center left part of the image. The specific function 142 selects representative slices from among three groups of medical images including the false detection candidate region, as shown in the upper right column, middle right column, and lower right column of FIG. 6, respectively.

Subsequently, the specific function 142 determines a group to be segmented from among the groups of medical images (step S111). Subsequently, the display control function 143 causes the display 130 to display the segmentation result of the representative slice selected by the specific function 142 and the false detection candidate area (step S113).

The display control function 143 causes the display 130 to display the segmentation result of the representative slice and the false detection candidate area, as well as the false detection candidate list and the input mode. The false detection candidate list indicates a list of false detection candidate groups. Among the multiple false positive candidate groups (the first false positive candidate group to the third false positive candidate group), the false positive candidate group (the first false positive candidate group) that is currently the target of weakly supervised data input is clear. The false detection candidate groups (the second false detection candidate group and the third false detection candidate group) that are displayed and are currently subject to input of weakly supervised data are darkly displayed. The input status of weakly supervised data in each false detection candidate group is also displayed.

The input mode is a mode that indicates the type of weakly supervised data that the user inputs through the input interface 120. The input modes include a background mode in which the user inputs to modify the background, and a target mode in which the user inputs to modify the segmentation target.

FIG. 7 is a diagram showing an example of a visualized image displayed on the display 130. A false detection candidate list is displayed on the upper left side of the display 130, and an input mode is displayed on the lower left side. On the right side of the display 130, a medical image GA10, a segmentation result GA20, and a false detection candidate area GA30 are displayed. Medical image GA10 is a representative slice in the first false detection candidate group.

Segmentation result GA20 is displayed superimposed on medical image GA10. In the example of FIG. 7, the segmentation result GA20 is displayed so as to cover the slightly upper left side of the medical image GA10. The false detection candidate area GA30 is located in the upper center of the medical image GA10 and near the right end of the segmentation result GA20. The user inputs weakly supervised data while viewing the image shown in FIG.

FIG. 8 is a diagram showing an example of a visualized image displayed on the display 130. The user operates the input interface 120 while viewing the image displayed on the display 130, thereby performing an input operation to instruct correction of the segmentation result GA20 within the false detection candidate area GA30.

For example, when the segmentation result GA20 is narrow, the user performs an input operation on the input interface 120 to specify an area to widen the segmentation result GA20 using the input mode. For example, if the segmentation result GA20 is too wide, the user operates the input interface 120 to specify an area to widen the background, using the input mode as the background. At this time, the user does not need to perform an input operation regarding the background or object for the entire erroneous detection candidate area, but only needs to perform an input operation for the range to be corrected.

Subsequently, the reception function 144 determines whether weakly supervised data corresponding to the correction of the segmentation result based on the user's input operation has been received (step S115). If the receiving function 144 determines that weakly supervised data is not accepted, it repeats the process of step S115 until weakly supervised data is accepted.

When the reception function 144 determines that the weakly supervised data has been accepted, the resegmentation function 145 determines the target based on the segmentation results of the medical images in the target false detection candidate group and the weakly supervised data received by the reception function 144. All medical images in the false detection candidate group are re-segmented (step S117).

FIG. 9 is a diagram illustrating an overview of the re-segmentation process. The columns from the left in FIG. 9 are referred to as first to fourth columns, respectively. The upper part of the first column shows a representative slice of the first false detection candidate group, the middle part of the first column shows a representative slice of the second false detection candidate group, and the lower part of the first column shows a representative slice of the third false detection candidate group.

With the representative slice of the first false detection candidate group shown in the upper part of the first column being displayed on the display 130, the reception function 144 receives weakly supervised data based on the user's input operation, thereby obtaining the segmentation result in the false detection candidate area. resegmentation is performed based on weakly supervised data.

Subsequently, re-segmentation is performed on all medical images of the first erroneous detection candidate group shown in the upper row of the third column based on the weakly supervised data. By performing the same process on the second false detection candidate group, the resegmentation in the false detection candidate area of the second false detection candidate group is weak, as shown in the middle row of the first column, the middle row of the second column, and the middle row of the third column. Executed based on supervised data.

Furthermore, by performing similar processing on the third false positive candidate group, resegmentation in the false positive candidate area of the second false positive candidate group is achieved, as shown in the lower row of the first column, the lower row of the second column, and the lower row of the third column. is performed based on weakly supervised data. By combining all the medical images shown in the upper row of the third column, the middle row of the third row, and the lower row of the third row, the segmentation results of all the medical images are obtained as shown in the fourth column.

Subsequently, the re-segmentation function 145 determines whether re-segmentation has been completed for all false detection candidate groups (step S119). If it is determined that re-segmentation has not been completed for all false detection candidate groups, the re-segmentation function 145 returns the process to step S111, determines the false detection candidate groups for which re-segmentation has not been completed, and continues the process up to step S119. Repeat the process.

When the re-segmentation function 145 determines that re-segmentation has been completed for all false detection candidate groups, the medical image processing apparatus 100 ends the process shown in FIG. 5.

The medical image processing apparatus 100 of the first embodiment displays a false detection candidate region and performs segmentation (re-segmentation) of the target region based on weakly supervised data input by the user who saw the false detection region. do. Therefore, the effort required for resegmenting medical images can be reduced.

(Second embodiment)
Next, a second embodiment will be described. The medical image processing apparatus 100 of the second embodiment differs from the medical image processing apparatus 100 of the first embodiment mainly in that it includes a model generation function. The medical image processing apparatus of the second embodiment will be described below, focusing on the differences from the first embodiment.

FIG. 10 is a block diagram showing an example of the configuration of a medical image processing apparatus 100 according to the second embodiment. The medical image processing apparatus 100 of the second embodiment includes a model generation function 210 in the processing circuit 140. The model generation function 210 updates the segmentation model 153 by machine learning using the re-segmentation results as input. The model generation function 210 is an example of a model generation unit.

The medical image processing apparatus 100 according to the second embodiment performs processing until segmentation (re-segmentation) of a target area in a medical image is performed based on the segmentation result and weakly supervised data using the same procedure as in the first embodiment. Execute. Subsequently, the model generation function 210 updates the segmentation model 153 by machine learning using the segmentation results and the supervised data with the weakly supervised data as input data and the re-segmentation results as output data. The model generation function 210 stores the updated segmentation model in the memory 150 as a new segmentation model 153.

The medical image processing apparatus 100 of the second embodiment has the same effects as the medical image processing apparatus 100 of the first embodiment. The medical image processing apparatus 100 of the second embodiment includes a model generation function 210. Therefore, each time re-segmentation is executed, the segmentation model 153 is updated using the re-segmentation results, so the accuracy of segmentation by the segmentation model 153 can be improved.

(Third embodiment)
Next, a third embodiment will be described. In the medical image processing apparatus 100 of the third embodiment, the segmentation result can be corrected multiple times at any timing or in real time. For this reason, the medical image processing apparatus 100 performs the first implementation in that when the re-segmentation function 145 executes re-segmentation based on weakly supervised data, the display control function 143 displays the re-segmentation result on the display 130. Mainly differs in form. The third embodiment will be described below, focusing on the differences from the first embodiment.

FIG. 11 is a diagram showing an example of a visualized image displayed on the display 130. In the medical image processing apparatus 100 of the third embodiment, the display control function 143 displays the medical image GA10, segmentation result GA20, false detection candidate area GA30, etc., as well as re-segmentation result GA21, continue button GA40, and completion button GA50. 130.

The re-segmentation result GA21 is obtained by re-segmentation by the re-segmentation function 145. Although the re-segmentation result GA21 is displayed separately from the segmentation result GA20, the segmentation result GA20 and the re-segmentation result GA21 may be displayed integrally. The continue button GA40 and the completion button GA50 will be explained in the fourth embodiment.

After the re-segmentation result GA21 is displayed, if re-segmentation is further performed by the re-segmentation function 145, the segmentation result GA20 here and the re-segmentation result GA21 are integrated and displayed on the display 130 as the segmentation result GA20. Is displayed.

The medical image processing apparatus 100 of the third embodiment has the same effects as the medical image processing apparatus 100 of the first embodiment. The medical image processing apparatus 100 of the third embodiment further displays a re-segmentation result GA21 when re-segmenting is performed again. Therefore, the user can modify the segmentation result at any timing or in real time while checking the re-segmentation result GA21.

(Fourth embodiment)
Next, a fourth embodiment will be described. The main feature of the medical image processing apparatus 100 of the fourth embodiment is that it is possible to correct the segmentation result multiple times at any timing or in real time, and then perform re-segmentation after re-segmenting the medical image. This is different from the first embodiment. The fourth embodiment will be described below, focusing on the differences from the first embodiment.

In the medical image processing apparatus 100 of the fourth embodiment, the re-segmentation results are displayed on the display 130 while re-segmentation is being performed. The re-segmentation results here include re-segmentation results from each re-segmentation when re-segmentation is performed multiple times. In the medical image processing apparatus 100 of the fourth embodiment, after performing re-segmentation, the user can select whether to further perform re-segmentation or to complete the segmentation.

Both the continuation button GA40 and the completion button GA50 shown in FIG. 11 are GUI. It can be input by the user specifying it using the input interface 120. By designating the continuation button GA40, continuation information indicating a continuation instruction is transmitted from the input interface 120 to the processing circuit 140. By designating the completion button GA50, completion information is transmitted from the input interface 120 to the processing circuit 140.

When the processing circuit 140 acquires the continuation information by the acquisition function 141, the reception function 144 accepts input of weakly supervised data from the user who has specified the input mode, and the re-segmentation function 145 executes re-segmentation. When re-segmentation is executed, the display control function 143 causes the display 130 to display the image shown in FIG. 11 including the re-segmentation results. When the processing circuit 140 acquires the completion information by the acquisition function 141, the processing circuit 140 ends the execution of the re-segmentation in the re-segmentation function 145.

The medical image processing apparatus 100 of the fourth embodiment has the same effects as the medical image processing apparatus 100 of the first embodiment. In the medical image processing apparatus 100 of the fourth embodiment, the display control function 143 further displays a continue button GA40 and a complete button GA50 on the display 130. Therefore, it is possible to reduce the effort required for the user to obtain the segmentation result that he/she desires.

In the medical image processing apparatus 100 of the fourth embodiment, the continue button GA40 and the complete button GA50 are displayed, but one or both of the continue button GA40 and the complete button GA50 may not be displayed. If the continue button GA40 is not displayed, for example, the re-segmentation function 145 may perform re-segmentation every time weakly supervised data is input. If the completion button GA50 is not displayed, the re-segmentation execution in the re-segmentation function 145 may be terminated, for example, when no teacher data is input for a certain period of time.

(Fifth embodiment)
Next, a fifth embodiment will be described. The main feature of the medical image processing apparatus 100 of the fifth embodiment is that the acquisition function 141 has a segmentation function that acquires a medical image transmitted by the modality 30 and segments the medical image acquired by the acquisition function 141. This is different from the first embodiment. The fifth embodiment will be described below, focusing on the differences from the first embodiment.

FIG. 12 is a block diagram showing an example of the configuration of a medical image processing apparatus 100 according to the fifth embodiment. The processing circuit 140 in the medical image processing apparatus 100 of the fifth embodiment includes a segmentation function 310. The segmentation function 310 automatically segments the medical image acquired by the acquisition function 141. The segmentation function 310 inputs a medical image and performs automatic segmentation. The segmentation function 310 stores the segmentation results in the automatic segmentation result DB 152. The segmentation function 310 is an example of a segmentation unit.

The medical image processing apparatus 100 of the fifth embodiment has the same effects as the medical image processing apparatus 100 of the first embodiment. The medical image processing apparatus 100 of the fifth embodiment further includes a segmentation function 310 that segments a medical image. Therefore, it is possible to reduce the effort required for the user to obtain segmentation results even for medical images on which segmentation has not been performed.

According to at least one embodiment described above, a medical image including a target region to be segmented, an acquisition unit that acquires information regarding a segmentation result of the target region included in the medical image, and an erroneous detection regarding the segmentation result. a specifying unit that identifies a candidate area; a display control unit that displays information related to the false detection candidate area in association with the medical image; a reception unit that receives labeling from a user regarding the false detection candidate area; By having a re-segmentation unit that performs segmentation of the target area, it is possible to reduce the effort required for re-segmenting a medical image.

Although several embodiments have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included within the scope and gist of the invention as well as within the scope of the invention described in the claims and its equivalents.

1...Hospital system 10...HIS
20...RIS
30...Modality 40...PACS
100...Medical image processing device 110...Communication interface 120...Input interface 130...Display 140...Processing circuit 141...Acquisition function 142...Specific function 143...Display control function 144...Reception function 145...Resegmentation function 150...Memory 151...Acquired image DB
152...Automatic segmentation result DB
153...Segmentation model 210...Model generation function 310...Segmentation function GA10...Medical image GA20...Segmentation result GA21...Resegmentation result GA30...False detection candidate area GA40...Continue button GA50...Complete button NW...Network

Claims (10)

an acquisition unit that acquires a medical image including a target region to be segmented, and information regarding a segmentation result of the target region included in the medical image;
a specifying unit that specifies a false detection candidate region regarding the segmentation result;
a display control unit that displays information regarding the false detection candidate region in association with the medical image;
a reception unit that receives labeling regarding the false detection candidate region from a user;
a resegmentation unit that performs resegmentation of the target area based on the labeling,
Medical image processing device. The acquisition unit acquires a plurality of the medical images,
The identifying unit selects a representative slice from among the plurality of medical images that are grouped based on a classification criterion;
The display control unit displays information regarding the false detection candidate region in association with the representative slice.
The medical image processing device according to claim 1. The identification unit groups the plurality of medical images using at least one of a distance or an image feature within the medical image as the distribution criterion.
The medical image processing device according to claim 2. The display control unit displays information regarding the false detection candidate area together with the segmentation result.
The medical image processing device according to claim 1. The specifying unit adjusts the number of representative slices according to the number of the plurality of medical images.
The medical image processing device according to claim 2. further comprising a model generation unit that generates a segmentation model by machine learning using the results of the re-segmentation as training data;
The medical image processing device according to claim 1. The display control unit displays the result of the re-segmentation together with information regarding the false detection candidate area.
The medical image processing device according to claim 1. The reception unit receives a continuation instruction to continue segmentation of the target area.
The medical image processing device according to claim 7. The computer is
obtaining a medical image including a target region to be segmented and information regarding a segmentation result of the target region included in the medical image;
identifying a false positive candidate region regarding the segmentation result;
displaying information regarding the false detection candidate region in association with the medical image;
receiving labeling regarding the false detection candidate area from a user;
performing segmentation of the region of interest based on the labeling;
Medical image processing method. to the computer,
obtaining a medical image including a target region to be segmented and information regarding a segmentation result of the target region included in the medical image;
identifying a false positive candidate region regarding the segmentation result;
displaying information regarding the false detection candidate region in association with the medical image;
receiving labeling regarding the false detection candidate area from the user;
performing segmentation of the region of interest based on the labeling;
program.

Images