Image processing device, method, and program

The image processing device enhances CAD systems by allowing operators to specify regions of interest, using trained models to accurately detect lesions in medical images based on indirect findings, aligning with physician interpretation.

JP2024168593APending Publication Date: 2024-12-05FUJIFILM CORP +1
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
JP2023085412
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-24
Publication Date
2024-12-05

AI Technical Summary

Technical Problem

Conventional computer-aided diagnosis (CAD) systems struggle to accurately detect lesions in medical images that are not clearly depicted, relying on direct visualization, while physicians often rely on indirect findings such as tissue properties and shape changes.

Method used

An image processing device and method that allows operators to specify regions of interest based on indirect findings, using trained models to enhance and extract abnormal regions from medical images, reflecting the operator's interpretation thoughts.

Benefits of technology

Accurately extracts abnormal regions from medical images by emphasizing latent expressions derived from specified regions of interest, improving detection accuracy in line with physician insights.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024168593000001_ABST
    Figure 2024168593000001_ABST
Patent Text Reader

Abstract

To accurately extract an abnormal region from a medical image by following a method according to the thought of an operator such as doctor interpreting the medical image in an image processing device, method and program.SOLUTION: A processor accepts an instruction to specify at least one region of interest that serves as the basis for an abnormal region included in a medical image and extracts the abnormal region from the medical image on the basis of the specified region of interest.SELECTED DRAWING: Figure 3
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to an image processing device, a method, and a program. [Background technology]

[0002] In recent years, advances in medical equipment such as CT (Computed Tomography) devices and MRI (Magnetic Resonance Imaging) devices have made it possible to perform image diagnosis using higher quality, higher resolution medical images. In addition, computer-aided diagnosis (CAD) has been put to practical use, which derives the probability of the presence of a lesion and its location information by analyzing medical images and presents this to doctors such as radiologists. For example, Patent Document 1 proposes a method of identifying an area of ​​a target organ and extracting an area suspected of being abnormal based on diagnostic criteria established for each organ. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Republished Publication No. 2007-119297 Summary of the Invention [Problem to be solved by the invention]

[0004] Depending on the type and size of the lesion or the imaging technique of the medical image, the lesion may not be clearly depicted in the medical image. For example, in contrast-enhanced tomographic images of the abdomen, tumors related to pancreatic cancer are relatively clearly depicted, but in non-contrast tomographic images, tumors related to pancreatic cancer are hardly depicted. Doctors may discover such barely depicted lesions by using indirect findings shown in medical images as a clue. Indirect findings indicate at least one of the characteristics of the properties and shape of the tissue surrounding the lesion, which appears as the lesion develops. Indirect findings include, for example, atrophy, swelling, and calcification.

[0005] Conventional CAD systems were developed on the assumption that lesions were clearly depicted in medical images to a certain extent, making it difficult to detect lesions that were barely visible as described above. For this reason, there was a demand for the development of a CAD system that would be able to detect lesions that were barely visible using indirect findings as a clue, in line with the thinking of doctors described above.

[0006] The present disclosure has been made in consideration of the above circumstances, and aims to accurately extract abnormal regions from medical images using a method that conforms to the thinking of an operator, such as a doctor, who interprets the medical images. [Means for solving the problem]

[0007] The image processing device according to the present disclosure includes at least one processor, The processor receives an instruction to identify at least one region of interest that is the basis of an abnormal region included in a medical image; An abnormal region is extracted from the medical image based on the specified region of interest.

[0008] In the image processing device according to the present disclosure, the processor extracts a region of interest from a medical image, Displaying a medical image so that a region of interest can be visually recognized; An instruction to specify a region of interest may be received by receiving an instruction to confirm or modify the region of interest.

[0009] In addition, in the image processing device according to the present disclosure, the processor extracts an abnormal region based on the medical image and the region of interest using the trained model, The trained model may be trained to emphasize the latent representations it derives based on a region of interest.

[0010] In the image processing device of the present disclosure, the processor displays a medical image, The instruction to identify a region of interest may be received by accepting designation of the region of interest in a displayed medical image.

[0011] The image processing method according to the present disclosure includes receiving an instruction to identify at least one region of interest that is a basis for an abnormal region included in a medical image; An abnormal region is extracted from the medical image based on the specified region of interest.

[0012] The image processing program according to the present disclosure includes a step of receiving an instruction to identify at least one region of interest that is a basis for an abnormal region included in a medical image; The computer is caused to execute a procedure of extracting an abnormal region from the medical image based on the specified region of interest. Effect of the Invention

[0013] According to the present disclosure, abnormal regions can be extracted from medical images with high accuracy. [Brief description of the drawings]

[0014] [Figure 1] FIG. 1 is a diagram showing a schematic configuration of a diagnosis support system to which an image processing device according to a first embodiment of the present disclosure is applied; [Diagram 2] FIG. 1 is a diagram showing a hardware configuration of an image processing apparatus according to a first embodiment; [Diagram 3] Functional configuration diagram of an image processing apparatus according to a first embodiment [Figure 4] FIG. 2 is a diagram illustrating a process performed in the image processing device according to the first embodiment; [Diagram 5] A schematic diagram showing the process performed by the derived model [Figure 6] A diagram showing the training data used to train the derived model. [Figure 7] Diagram for explaining the learning of the derived model [Figure 8] A diagram showing the display screen of the extraction results [Figure 9] 1 is a flowchart showing the processing performed in the first embodiment. [Figure 10] Functional configuration diagram of an image processing apparatus according to a second embodiment [Figure 11]1 is a flowchart showing the process performed in the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Hereinafter, an embodiment of the present disclosure will be described with reference to the drawings. First, the configuration of a medical information system to which an image processing device according to this embodiment is applied will be described. Fig. 1 is a diagram showing a schematic configuration of a medical information system. In the medical information system shown in Fig. 1, a computer 1 including an image processing device according to this embodiment, an imaging device 2, and an image storage server 3 are connected in a communicable state via a network 4.

[0016] The computer 1 includes an image processing device according to this embodiment, and has the image processing program of this embodiment installed. The computer 1 may be a workstation or personal computer operated directly by a doctor who performs a diagnosis, or may be a server computer connected to them via a network. The image processing program is stored in a storage device of a server computer connected to the network, or in a network storage, in a state accessible from the outside, and is downloaded to the computer 1 used by the doctor upon request and installed. Alternatively, the program is distributed by recording on a recording medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory), and is installed in the computer 1 from the recording medium.

[0017] The imaging device 2 is a device that generates a three-dimensional image representing a diagnostic target part of a subject by imaging the part, specifically, a CT device, an MRI device, a PET (Positron Emission Tomography) device, etc. The three-dimensional image composed of a plurality of tomographic images generated by the imaging device 2 is transmitted to the image storage server 3 and stored. In this embodiment, the imaging device 2 is a CT device, and generates a CT image of the abdomen of the subject as a three-dimensional image. The acquired CT image may be a contrast-enhanced CT image or a non-contrast CT image.

[0018] The image storage server 3 is a computer that stores and manages various data, and is equipped with a large-capacity external storage device and database management software. The image storage server 3 communicates with other devices via a wired or wireless network 4, and transmits and receives image data and the like. Specifically, it acquires various data, including image data of CT images generated by the imaging device 2, via the network, and stores and manages them in a recording medium such as a large-capacity external storage device. The storage format of the image data and the communication between the devices via the network 4 are based on protocols such as DICOM (Digital Imaging and Communication in Medicine).

[0019] Next, an image processing device according to the first embodiment will be described. FIG. 2 is a diagram showing a hardware configuration of the image processing device according to the first embodiment. As shown in FIG. 2, the image processing device 20 includes a CPU (Central Processing Unit) 11, a non-volatile storage 13, and a memory 16 as a temporary storage area. The image processing device 20 also includes a display 14 such as a liquid crystal display, an input device 15 such as a keyboard and a mouse, and a network I / F (InterFace) 17 connected to a network 4. The CPU 11, the storage 13, the display 14, the input device 15, the memory 16, and the network I / F 17 are connected to a bus 18. The CPU 11 is an example of a processor in the present disclosure.

[0020] The storage 13 is realized by a hard disk drive (HDD), a solid state drive (SSD), a flash memory, etc. The storage 13 as a storage medium stores an image processing program 12. The CPU 11 reads the image processing program 12 from the storage 13, expands it in the memory 16, and executes the expanded image processing program 12.

[0021] Next, the functional configuration of the image processing device according to the first embodiment will be described. Fig. 3 is a diagram showing the functional configuration of the image processing device according to the first embodiment, and Fig. 4 is a diagram showing the process performed by the image processing device according to the first embodiment. As shown in Fig. 3, the image processing device 20 includes an image acquisition unit 21, a first extraction unit 22, a second extraction unit 23, a display control unit 24, and a third extraction unit 25. Then, the CPU 11 executes the image processing program 12, so that the CPU 11 functions as the image acquisition unit 21, the first extraction unit 22, the second extraction unit 23, the display control unit 24, and the third extraction unit 25.

[0022] The image acquisition unit 21 acquires a medical image G0 to be processed from the image storage server 3 in response to an instruction from an operator via the input device 15. In this embodiment, the medical image G0 is a CT image consisting of multiple tomographic images including the abdomen of a human body.

[0023] The first extraction unit 22 extracts a region of a target organ from the medical image G0. In this embodiment, the target organ is the pancreas. For this reason, the first extraction unit 22 has a semantic segmentation model (hereinafter, referred to as an SS (Semantic Segmentation) model) 22A that has been machine-learned to extract the pancreas from the medical image G0. As is well known, the SS model 22A is a machine-learning model that outputs an output image in which a label representing an extraction target (class) is assigned to each pixel of an input image. In this embodiment, the input image is a tomographic image that constitutes the medical image G0, the extraction target is the pancreas, and the output image is an image in which the region of the pancreas is labeled. The SS model is constructed by a convolutional neural network (CNN) such as Residual Networks (ResNet) and U-shaped Networks (U-Net).

[0024] The first extraction unit 22 uses such an SS model to extract a region 30 of the pancreas included in the medical image G0, as shown in Fig. 4. The extraction of the target organ is not limited to using the SS model. Any method for extracting the target organ from the medical image G0, such as template matching or threshold processing for the CT value, can be applied.

[0025] The second extraction unit 23 extracts at least one region of interest that is the basis of an abnormal region in the pancreas included in the medical image G0. In this embodiment, a region of pancreatic duct stenosis in the pancreas is extracted from the medical image G0.

[0026] Here, when a doctor interprets the medical image G0 to identify a lesion in the pancreas, the doctor uses indirect findings that indicate changes in the properties and shape of the tissue surrounding the lesion, such as pancreatic duct stenosis, as a clue. The "indirect" in "indirect findings" is used in contrast to "direct" findings that directly link lesions, such as tumors, to diseases, such as cancer. In addition, the doctor uses the pancreatic duct in the pancreas as a clue to identify pancreatic duct stenosis.

[0027] In the first embodiment, the second extraction unit 23 extracts a region of pancreatic duct stenosis from the medical image G0 as an indirect finding of an abnormality in the pancreas. For this purpose, the second extraction unit 23 has an SS model 23A that has been machine-trained to extract a region of pancreatic duct stenosis from the medical image G0 based on the medical image G0 and a region 30 of the pancreas included in the medical image G0. The second extraction unit 23 uses such SS model 23A to extract a region 31 of pancreatic duct stenosis included in the region 30 of the pancreas included in the medical image G0, as shown in FIG.

[0028] The display control unit 24 displays the medical image G0 from which the region 31 of pancreatic duct stenosis has been extracted on the display 14. Note that the medical image G0 displayed on the display 14 has a mask applied to the region 31 of pancreatic duct stenosis as shown in FIG.

[0029] The operator looks at the medical image G0 displayed on the display 14 and uses the input device 15 to specify the region of interest, i.e., the region 31 of pancreatic duct stenosis. Here, the specifying instruction includes an instruction to confirm the extracted region of interest or an instruction to correct the extracted region of interest. That is, if the region 31 of pancreatic duct stenosis included in the displayed medical image G0 correctly represents the region of pancreatic duct stenosis, the operator specifies the region of interest by issuing an instruction to confirm from the input device 15. On the other hand, if the region 31 of pancreatic duct stenosis included in the displayed medical image G0 does not correctly represent the region of pancreatic duct stenosis, the operator specifies the region of interest by issuing an instruction to correct the region of interest from the input device 15. Note that G1 is used as the reference symbol for the medical image for which the specifying instruction has been issued.

[0030] The third extraction unit 25 receives an instruction to specify a region of interest, and extracts an abnormal region from the medical image G0 based on the region of interest specified in the instruction. That is, the third extraction unit 25 extracts a pancreatic lesion region 32 as an abnormal region from the medical image G0 based on the region of pancreatic duct stenosis specified in the instruction, and derives an output image G2 from which the abnormal region is extracted. For this purpose, the third extraction unit 25 has an extraction model 25A that has been machine-trained to extract a pancreatic lesion region from the medical image G0 and the region of interest (i.e., the region of pancreatic duct stenosis).

[0031] FIG. 5 is a diagram showing a schematic diagram of the process performed by the extraction model 25A. As shown in FIG. 5, the extraction model 25A has an encoder 41, an attention 42, and a decoder 43. When a medical image G0 is input, the encoder 41 outputs a latent representation z1 in which the features of the medical image G0 are dimensionally compressed in order to extract a lesion area of ​​the pancreas. The extraction model 25A extracts a lesion area of ​​the pancreas from the medical image G0. Therefore, the latent representation z1 includes features of a region of interest for extracting a lesion area of ​​the pancreas.

[0032] Here, when a doctor interprets the medical image G0 to find a lesion in the pancreas, the region of pancreatic duct stenosis in the pancreas is used as a clue. Therefore, the region of interest for extracting the region of the lesion in the pancreas is the region of pancreatic duct stenosis. FIG. 5 shows an image zg1 in which the region of interest for extracting the region of the lesion in the pancreas derived in the encoder 41 is visualized. In the image zg1, a region near the pancreatic duct stenosis in the medical image G0 is extracted as the region of interest A1. Note that the region of interest A1 may be extracted by masking the region of interest A1, or may be extracted by surrounding the region of interest A1 with a bounding box.

[0033] The attention 42 in the extraction model 25A derives a corrected latent expression z1' by correcting the latent expression z1 derived by the encoder 41 based on the region of interest extracted by the second extraction unit 23 and specified by the operator. Specifically, the corrected latent expression z1' is derived by emphasizing the region of pancreatic duct stenosis in the latent expression z1. For example, the emphasis method may be a method of multiplying the region of pancreatic duct stenosis in the latent expression z1 by a weight greater than 1 that is determined in advance based on the region of interest specified by the operator. FIG. 5 shows an image zg1' in which the corrected latent expression z1' is visualized. In the image zg1', the region of pancreatic duct stenosis in the medical image G0 is shown as an emphasized region A1'.

[0034] The decoder 43 extracts the pancreatic lesion region in the medical image G0 by reconstructing the corrected latent representation z1', and derives an output image G2 in which the pancreatic lesion region 32 is labeled.

[0035] Next, the learning of the extraction model 25A will be described. Fig. 6 is a diagram showing training data used in the training of the extraction model 25A, and Fig. 7 is a diagram for explaining the training of the extraction model 25A. As shown in Fig. 6, the training data 50 includes a training medical image 51 and correct answer data 52. The training medical image 51 is a medical image including a pancreas. The correct answer data 52 includes a first mask image 53 in which a region of pancreatic duct stenosis in the training medical image 51 is masked, and a second mask image 54 in which a lesion region of the pancreas in the training medical image 51 is masked.

[0036] 7, a training medical image 51 is first input to the encoder 41, and a latent representation z01 in which the features of the training medical image 51 are dimensionally compressed for extracting a lesion area of ​​the pancreas is output from the encoder 41. The extraction model 25A extracts the lesion area of ​​the pancreas by focusing on the area of ​​pancreatic duct stenosis. For this reason, an image 60 is derived that visualizes the area of ​​interest, i.e., the area of ​​pancreatic duct stenosis, which is the basis for extracting the lesion area of ​​the pancreas derived in the encoder 41, and the difference between the image 60 and the first mask image 53, which is the correct answer data 52, is derived as a loss L1.

[0037] The image 60 may be derived by configuring the encoder 41 to have an output layer for visualizing a region of interest for extracting a derived lesion region of the pancreas. Alternatively, the encoder 41 may not be provided with an output layer for visualizing a region that serves as a basis for extracting a lesion region of the pancreas, and the region of interest for extracting a lesion region of the pancreas may be visualized using the Grad-CAM technique.

[0038] Further, the latent representation z01 output from the encoder 41 is corrected by the attention 42 so as to emphasize the region of the pancreatic duct stenosis based on the first mask image 53, and the corrected latent representation z01' is derived. Then, the corrected latent representation z01' is reconstructed by the decoder 43 to derive a feature amount representing the region of the pancreatic lesion in the training medical image 51, and an output image 61 is derived in which the region in which the feature amount is equal to or greater than a predetermined threshold is masked. The output image 61 is an image derived by the extraction model 25A in which the region of the pancreatic lesion in the training medical image 51 is masked. Then, the difference between the output image 61 and the second mask image 54 is derived as the loss L2.

[0039] Then, the encoder 41, the attention 42, and the decoder 43 are trained so that the losses L1 and L2 are small. The attention 42 is trained using the loss L1 so that the area of ​​pancreatic duct stenosis is emphasized by the attention 42. As a result, the encoder 41 is trained to derive a latent representation representing the characteristics of pancreatic duct stenosis. The attention 42 is trained so that the area of ​​pancreatic duct stenosis in the pancreas is emphasized. The decoder 43 is trained to extract the area of ​​the pancreatic lesion from the latent representation corrected by the attention 42. Then, the extraction model 25A is constructed by repeating the training using multiple teaching data until the losses L1 and L2 become equal to or less than a predetermined threshold value, or by repeating the training a predetermined number of times.

[0040] The display control unit 24 displays the output image G2 in which the abnormal region of the pancreas extracted by the third extraction unit 25 is masked on the display 14. Fig. 8 is a diagram showing a display screen of the output image. As shown in Fig. 8, the output image G2 in which the lesion region of the pancreas derived by the third extraction unit 25 is masked is displayed on the display screen 70.

[0041] Next, the process performed in the first embodiment will be described. FIG. 9 is a flow chart showing the process performed in the first embodiment. First, the image acquisition unit 21 acquires the medical image G0 from the storage 13 (step ST1), and the first extraction unit 22 extracts the pancreas region 30, which is the region of the target organ, from the medical image G0 (step ST2). Next, the second extraction unit 23 extracts the region 31 of pancreatic duct stenosis, which is the region of interest that is the basis of the abnormal region included in the medical image G0 (step ST3). Then, the display control unit 24 displays the medical image G0 from which the region of interest has been extracted on the display 14 (step ST4).

[0042] The third extraction unit 25 then receives an instruction to specify the region of interest (step ST5), and extracts the pancreatic lesion region 32, which is an abnormal region, from the medical image G0 based on the region of interest for which the instruction has been given (step ST6).Then, the display control unit 24 displays the output image G2 in which the extracted abnormal region is masked on the display 14 (step ST7), and the process ends.

[0043] In this manner, in this embodiment, an instruction to specify a region of interest that is the basis of an abnormal region included in the medical image G0 is received, and an abnormal region is extracted from the medical image G0 based on the region of interest specified in the instruction. Therefore, according to this embodiment, the thinking of an operator such as a doctor can be reflected in the region of interest, and as a result, an abnormal region included in the medical image G0 can be detected with high accuracy according to a method that conforms to the thinking of an operator such as a doctor who interprets the medical image G0.

[0044] Furthermore, by emphasizing the latent representation derived by the extraction model 25A provided in the third extraction unit 25 based on the region of interest, abnormal regions can be detected more accurately from the medical image G0 according to a method that is more in line with the thinking of an operator, such as a doctor, who interprets the medical image G0.

[0045] In the first embodiment, after the second extraction unit 23 extracts the region of interest, the operator issues a confirmation or correction instruction to specify the region of interest. However, the present invention is not limited to this. A medical image G0 from which the region 30 of the pancreas is extracted may be displayed, and only the operator's instruction to specify the region of interest may be accepted. This will be described below as a second embodiment.

[0046] Fig. 10 is a diagram showing the functional configuration of an image processing device according to a second embodiment. In Fig. 10, the same components as those in Fig. 3 are given the same reference numerals, and detailed descriptions thereof will be omitted. The image processing device 20A according to the second embodiment differs from the first embodiment in that it does not include a second extraction unit 23.

[0047] Next, the process performed in the second embodiment will be described. Fig. 11 is a flow chart showing the process performed in the second embodiment. First, the image acquisition unit 21 acquires the medical image G0 from the storage 13 (step ST11), and the first extraction unit 22 extracts the pancreas region 30, which is the region of the target organ, from the medical image G0 (step ST12). Next, the display control unit 24 displays the medical image G0 from which the pancreas region 30 has been extracted on the display 14 (step ST13).

[0048] The third extraction unit 25 then receives an instruction to specify the region of interest (step ST14), and extracts the pancreatic lesion region 32, which is an abnormal region, from the medical image G0 based on the region of interest for which the instruction has been given (step ST15).Then, the display control unit 24 displays the output image G2 in which the extracted abnormal region is masked on the display 14 (step ST16), and the process ends.

[0049] As a result, according to the second embodiment as well, an abnormal region included in the medical image G0 can be detected with high accuracy in accordance with a method that conforms to the thinking of an operator, such as a doctor, who interprets the medical image G0.

[0050] In the above-mentioned embodiments, the instruction to specify only the region of pancreatic duct stenosis of the pancreas as the region of interest is received, but the present invention is not limited to this. An instruction to specify a plurality of regions of interest may be received. For example, an instruction to specify a region of pancreatic swelling in addition to the region of pancreatic duct stenosis as the region of interest may be received, and the third extraction unit 25 may extract an abnormal region from the medical image G0 based on the region of pancreatic duct stenosis and the region of pancreatic swelling.

[0051] In addition, in each of the above embodiments, the target organ is the pancreas, but is not limited thereto. In addition to the pancreas, any organ such as the brain, heart, lungs, and liver can be used as the target organ.

[0052] In the above embodiment, a CT image is used as the medical image G0, but this is not limited to this. In addition to three-dimensional images such as MRI images, any image such as a radiological image obtained by simple imaging can be used as the medical image G0.

[0053] In the above embodiment, the following various processors can be used as the hardware structure of the processing unit that executes various processes, such as the image acquisition unit 21, the first extraction unit 22, the second extraction unit 23, the display control unit 24, and the third extraction unit 25. As described above, the above various processors include a CPU, which is a general-purpose processor that executes software (programs) and functions as various processing units, as well as a programmable logic device (PLD), which is a processor whose circuit configuration can be changed after manufacture, such as an FPGA (Field Programmable Gate Array), and a dedicated electric circuit, such as an ASIC (Application Specific Integrated Circuit), which is a processor having a circuit configuration designed specifically for executing specific processes.

[0054] A single processing unit may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (e.g., a combination of multiple FPGAs or a combination of a CPU and an FPGA). Also, multiple processing units may be configured with a single processor.

[0055] As an example of configuring multiple processing units with one processor, first, there is a form in which one processor is configured with a combination of one or more CPUs and software, as typified by computers such as client and server, and this processor functions as multiple processing units. Second, there is a form in which a processor is used that realizes the functions of the entire system including multiple processing units with one IC (Integrated Circuit) chip, as typified by System On Chip (SoC), etc. In this way, the various processing units are configured using one or more of the above various processors as a hardware structure.

[0056] Furthermore, more specifically, the hardware structure of these various processors can be an electric circuit that combines circuit elements such as semiconductor elements. [Explanation of symbols]

[0057] 1. Computer 2. Imaging Device 3. Image storage server 4. Network 11 CPU 12 Image Processing Programs 13. Storage 14 Display 15 Input Devices 16 Memory 17 Network Interface 18 Bus 20 Image processing device 21 Image acquisition unit 22 1st extraction part 22A SS model 23 Second extraction part 23A SS model 24 Display control section 25 Third extraction part 25A Extraction Model 30 Pancreatic Region 31 Area of ​​pancreatic duct stenosis 32 Lesion area 41 Encoder 42 Attention 43 Decoder 50 Teacher Data 51 Educational Medical Images 52 Correct data 53 First mask image 54 Second mask image 60 images 61 Output images 70 display screen G0,G1 Medical Images G2 Output Image L1, L2 loss z1,z1′,z01,z01′ latent representation Visualization of the latent representations zg1 and zg1′

Claims

1. At least one processor; The processor, receiving an instruction to identify at least one region of interest that is the basis of an abnormal region included in a medical image; An image processing device that extracts the abnormal region from the medical image based on the specified region of interest.

2. The processor extracts the region of interest from the medical image; displaying the medical image so that the region of interest is visible; The image processing apparatus according to claim 1 , further comprising: an instruction to specify the region of interest by receiving an instruction to confirm or modify the region of interest.

3. The processor extracts the abnormal region based on the medical image and the region of interest using a trained model; The image processing device according to claim 2 , wherein the trained model is trained to emphasize a latent expression derived by the trained model based on the region of interest.

4. The processor displays the medical image; The image processing apparatus according to claim 1 , further comprising: a step of receiving an instruction to specify the region of interest by receiving designation of the region of interest in the displayed medical image.

5. receiving an instruction to identify at least one region of interest that is the basis of an abnormal region included in a medical image; An image processing method for extracting the abnormal region from the medical image based on the specified region of interest.

6. A step of receiving an instruction to identify at least one region of interest that is the basis of an abnormal region included in a medical image; and extracting the abnormal area from the medical image based on the specified area of ​​interest.

Citation Information

Patent Citations

  • Computer-aided detection system and method

    JP2008541889A

  • Diagnostic assistance system, diagnostic assistance method, and program

    JP2019149093A

  • Image processing device, method, and program

    WO2022270151A1