A medical image segmentation method, device, equipment and storage medium

Abstract

Embodiments of the present application disclose a medical image segmentation method, device and equipment and a storage medium, wherein the method comprises: acquiring a medical image to be segmented; inputting the medical image to be segmented into a target medical image segmentation model to obtain a segmentation result of the medical image to be segmented; wherein the target medical image segmentation model is a model trained in a generative adversarial manner based on a preset source domain medical sample image and a preset target domain medical sample image, wherein the preset target domain medical sample image is a medical sample image having the same region to be segmented as the preset source domain medical image and at least one different image feature. The technical solution of the embodiments of the present application solves the problem of insufficient accuracy of the medical image segmentation model trained by the prior art when performing medical image segmentation, and can train the medical image segmentation model in a generative adversarial manner to improve the segmentation accuracy of the medical image segmentation model.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a medical image segmentation method, apparatus, device and storage medium. Background Technology

[0002] Medical image segmentation is a crucial part of medical analysis, involving the segmentation of regions of interest within medical images. However, images from different hospitals exhibit varying data distributions and styles, leading to lower performance of segmentation models when applied in general applications. This discrepancy between the source and target domain data distributions is known as the domain offset problem, which results in insufficient accuracy for trained medical image segmentation models. Summary of the Invention

[0003] This invention provides a medical image segmentation method, apparatus, device, and storage medium, which can improve the segmentation accuracy of medical image segmentation models.

[0004] In a first aspect, embodiments of the present invention provide a medical image segmentation method, the method comprising:

[0005] Acquire the medical image to be segmented;

[0006] The medical image to be segmented is input into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented;

[0007] The target medical image segmentation model is a model trained by generating adversarial methods based on a preset source domain medical sample image and a preset target domain medical sample image. The preset target domain medical sample image is a medical sample image that has the same region to be segmented as the preset source domain medical image and has at least one different image feature.

[0008] In a second aspect, embodiments of the present invention provide a medical image segmentation apparatus, the apparatus comprising:

[0009] The medical image acquisition module is used to acquire the medical image to be segmented.

[0010] The medical image segmentation module is used to input the medical image to be segmented into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented;

[0011] The target medical image segmentation model is a model trained by generating adversarial methods based on a preset source domain medical sample image and a preset target domain medical sample image. The preset target domain medical sample image is a medical sample image that has the same region to be segmented as the preset source domain medical image and has at least one different image feature.

[0012] Thirdly, embodiments of the present invention provide a computer device, the computer device comprising:

[0013] One or more processors;

[0014] Memory, used to store one or more programs;

[0015] When the one or more programs are executed by the one or more processors, the one or more processors implement the medical image segmentation method described in any embodiment.

[0016] Fourthly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the medical image segmentation method described in any embodiment.

[0017] The technical solution provided by this invention obtains the segmentation result of the medical image to be segmented by inputting the medical image to be segmented into a target medical image segmentation model. The target medical image segmentation model is a model trained using a generative adversarial approach based on a preset source domain medical sample image and a preset target domain medical sample image. The preset target domain medical sample image is a medical sample image that has the same region to be segmented as the preset source domain medical image but possesses at least one different image feature. This invention solves the problem of insufficient accuracy in medical image segmentation models trained in the prior art. It improves the segmentation accuracy of medical image segmentation models by training them using a generative adversarial approach. Attached Figure Description

[0018] Figure 1 This is a flowchart of a medical image segmentation method provided in an embodiment of the present invention;

[0019] Figure 2 This is a flowchart of a medical image segmentation method provided in an embodiment of the present invention;

[0020] Figure 3 This is a flowchart illustrating the training process of a medical image segmentation model provided in an embodiment of the present invention.

[0021] Figure 4 This is a schematic diagram of the structure of a medical image segmentation device provided in an embodiment of the present invention;

[0022] Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Figure 1 This is a flowchart of a medical image segmentation method provided by an embodiment of the present invention. The embodiment of the present invention can be applied to scenarios where medical images are segmented. The method can be executed by a medical image segmentation device, which can be implemented by software and / or hardware.

[0025] like Figure 1 As shown, the medical image segmentation method includes the following steps:

[0026] S110. Obtain the medical image to be segmented.

[0027] The medical image to be segmented can be a medical image that requires image segmentation. For example, the medical image to be segmented can be an X-ray image of a certain part of a patient. In this embodiment of the invention, the medical image to be segmented needs to be analyzed to segment out the region of interest in the medical image.

[0028] S120. Input the medical image to be segmented into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented.

[0029] The target medical image segmentation model can be a model trained by generating adversarial methods based on a preset source domain medical sample image and a preset target domain medical sample image. The preset target domain medical sample image is a medical sample image that has the same region to be segmented as the preset source domain medical image and has at least one different image feature.

[0030] Specifically, preset source domain medical sample images and preset target domain medical sample images can be input into the initial medical image segmentation model, and the encoder of the initial medical image segmentation model can extract features from the sample image pairs to obtain source domain image features and target domain image features. The source domain image features and target domain image features are then input into the preset image domain recognizer and the decoder of the initial medical image segmentation model to obtain the corresponding image domain recognition results and image segmentation results, respectively. Finally, the parameters of the initial medical image segmentation model are adjusted based on the image domain recognition results and image segmentation results to obtain the target medical image segmentation model.

[0031] Specifically, the encoder in the initial medical image segmentation model can be adjusted in reverse based on the image domain recognition results, so that the preset image domain recognizer cannot identify whether the input medical image is a preset source domain medical sample image or a preset target domain medical sample image, thereby realizing a generative adversarial training method. Training the initial medical image segmentation model through generative adversarial training to obtain the target medical image segmentation model can improve the generalization ability and segmentation accuracy of the target medical image segmentation model, that is, improve the accuracy of the segmentation results of the medical image to be segmented.

[0032] The technical solution provided by this invention obtains the segmentation result of the medical image to be segmented by inputting the medical image to be segmented into a target medical image segmentation model. The target medical image segmentation model is a model trained through a generative adversarial approach based on preset source domain medical sample images and preset target domain medical sample images. The preset target domain medical sample images are medical sample images that have the same region to be segmented as the preset source domain medical images but possess at least one different image feature. This invention solves the problem of insufficient accuracy in medical image segmentation models trained by existing technologies. It improves the segmentation accuracy of medical image segmentation models by training them through a generative adversarial approach.

[0033] Figure 2 This is a flowchart of a medical image segmentation method provided by an embodiment of the present invention. This embodiment is applicable to scenarios involving the segmentation of medical images. Based on the above embodiments, this embodiment further explains how to train a target medical image segmentation model using a generative adversarial approach, based on preset source domain medical sample images and preset target domain medical sample images, and how to perform medical image segmentation based on the target medical image segmentation model. This device can be implemented in software and / or hardware and integrated into a computer device with application development capabilities.

[0034] like Figure 2 As shown, the medical image segmentation method includes the following steps:

[0035] S210. Take a preset source domain medical sample image with sample labels and a preset target domain medical sample image without labels as a set of sample image pairs.

[0036] The preset source domain medical sample image can be a medical sample image with sample labels, while the target domain medical sample image can be a medical sample image without sample labels. The preset target domain medical sample image and the preset source domain medical image have the same region to be segmented and at least one different image feature. For example, the preset target domain medical sample image and the preset source domain medical image can be chest X-ray images of different patients. A preset source domain medical sample image and a preset target domain medical sample image are used as a sample image pair, which can be used as a set of samples for training the subsequent medical image segmentation model.

[0037] S220. Input the sample image pairs into the initial medical image segmentation model respectively, and extract the source domain image features and target domain image features from the sample image pairs through the encoder of the initial medical image segmentation model respectively.

[0038] The initial medical image segmentation model can be an untrained, raw medical image segmentation model. Specifically, the initial medical image segmentation model includes a U-Net network, which comprises an encoder and a decoder. The encoder is used to extract features from the medical image, and the decoder is used to upsample the features to reconstruct the segmented image. The source domain image features are the image features of the preset source domain medical sample images, and the target domain image features are the image features of the preset target domain medical sample images. After inputting the sample image pairs into the initial medical image segmentation model, the encoder can extract features from the preset source domain medical sample images and the preset target domain medical sample images, respectively, to obtain the source domain image features and the target domain image features.

[0039] S230. Input the source domain image features and the target domain image features into the decoder of the preset image domain recognizer and the initial medical image segmentation model to obtain the corresponding image domain recognition results and image segmentation results, respectively.

[0040] The preset image domain recognizer can be a preset recognizer used to identify preset source domain medical sample images and preset target domain medical sample images. That is, the preset image domain recognizer can analyze the features of the input medical image to identify whether the medical image corresponding to the features is a preset source domain medical sample image or a preset target domain medical sample image. The recognition result output by the preset image domain recognizer is the image domain recognition result. The decoder of the initial medical image segmentation model can analyze the image features to segment the features of the region of interest, obtaining the image segmentation result. The image segmentation result includes the image domain recognition result of the preset target domain medical sample image and the image domain recognition result of the preset source domain medical sample image.

[0041] S240. Adjust the parameters of the initial medical image segmentation model based on the image domain recognition result and the image segmentation result to obtain the target medical image segmentation model.

[0042] The target medical image segmentation model can be a model obtained by training the initial medical image segmentation model. The parameters of the initial medical image segmentation model can be adjusted based on the image domain recognition results and image segmentation results to obtain the target medical image segmentation model.

[0043] Specifically, the image domain recognition loss function can be determined based on the image domain recognition results of the preset target domain medical sample image and the image domain recognition results of the preset source domain medical sample image; the target domain image segmentation loss function can be determined based on the image segmentation results of the preset target domain medical sample image; the source domain image segmentation loss function can be determined based on the image segmentation results of the preset source domain medical sample image; and the parameter values ​​of the encoder and decoder can be adjusted based on the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function to obtain the target medical image segmentation model.

[0044] The process of determining the target domain image segmentation loss function based on the image segmentation results of the medical sample image in the preset target domain includes: calculating the image pixel entropy of each pixel in the medical sample image in the preset target domain based on the image segmentation results; and determining the target domain image segmentation loss function based on the image pixel entropy. Specifically, the formula for calculating the image pixel entropy is: in Let Px be the entropy of the image pixel with coordinates (h, w), which is also the image segmentation result for that pixel; the formula for the target domain image segmentation loss function is: Where L ent (x t Let represent the target domain image segmentation loss function. The process of determining the source domain image segmentation loss function based on the image segmentation results of the preset source domain medical sample images includes: calculating the cross-entropy loss by combining the image segmentation results of the preset source domain medical sample images with the sample labels of the images to obtain the source domain image segmentation loss function.

[0045] For example, Figure 3 This is a flowchart illustrating the training process of a medical image segmentation model provided in an embodiment of the present invention, such as... Figure 3As shown, the medical image segmentation model includes an encoder and a decoder. The training process of the medical image segmentation model is as follows: Labeled source domain images and unlabeled target domain images are input into the encoder for feature extraction to obtain source domain image features and target domain image features; then, the source domain image features and target domain image features are input into the domain classifier and decoder to obtain the corresponding image domain recognition results and image segmentation results, respectively; finally, the image domain recognition loss function L is determined based on the image domain recognition results. dis The target domain image segmentation loss function L is determined based on the image segmentation results of the medical sample images in the preset target domain. ent (x t Based on the image segmentation results of the preset source domain medical sample images, the source domain image segmentation loss function L is determined. seg Finally, the parameter values ​​of the medical image segmentation model are adjusted according to the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function to complete the training process of the medical image segmentation model.

[0046] Furthermore, since the initial medical image segmentation model needs to be trained through generative adversarial methods, it is necessary to ensure that the preset image domain recognizer cannot distinguish between preset source domain medical sample images and preset target domain medical sample images. That is, the encoder and the preset image domain recognizer are used as the simulator and discriminator in adversarial training, respectively. The parameter values ​​of the encoder are adjusted based on the image domain recognition results and image segmentation results, so that the preset image domain recognizer cannot distinguish whether the input medical image is a preset source domain medical sample image or a preset target domain medical sample image.

[0047] Specifically, the encoder loss function can be obtained by summing the inverses of the target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function. The encoder parameter adjustment values ​​are then determined based on the encoder loss function, and the encoder parameter values ​​are adjusted accordingly. Specifically, by summing the inverses of the target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function when determining the encoder loss function, the encoder parameter values ​​can be adjusted based on the inverse of the image domain recognition loss function. This allows the adjusted encoder to extract image features that prevent the preset image domain recognizer from correctly recognizing the domain, thus achieving a generative adversarial training method. Training the initial medical image segmentation model using this generative adversarial training method to obtain the target medical image segmentation model can improve the generalization ability and segmentation accuracy of the target medical image segmentation model.

[0048] Correspondingly, while adjusting the encoder's parameter values, it is also necessary to adjust the decoder's parameter values. Specifically, the target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function can be summed to obtain the decoder loss function; the decoder parameter adjustment value is determined based on the decoder loss function, and the decoder parameter values ​​are adjusted according to the decoder parameter adjustment value.

[0049] S250. Obtain the medical image to be segmented, and input the medical image to be segmented into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented.

[0050] The medical image to be segmented can be any medical image that requires image segmentation. Inputting the medical image to be segmented into the target medical image segmentation model trained as described above will yield the segmentation result. Training the initial medical image segmentation model using a generative adversarial training method to obtain the target medical image segmentation model can improve the generalization ability and segmentation accuracy of the target medical image segmentation model, and consequently improve the segmentation accuracy of the medical image to be segmented.

[0051] The technical solution provided by this invention involves treating a pre-defined source domain medical sample image with sample labels and a pre-defined target domain medical sample image without labels as a pair of sample images. These sample image pairs are input into an initial medical image segmentation model, and the encoder of the initial model extracts features from each pair to obtain source domain image features and target domain image features. The source domain image features and target domain image features are then input into a pre-defined image domain recognizer and a decoder of the initial medical image segmentation model to obtain corresponding image domain recognition and image segmentation results, respectively. Based on the image domain recognition and segmentation results, the parameters of the initial medical image segmentation model are adjusted to obtain a target medical image segmentation model. Finally, the medical image to be segmented is acquired and input into the target medical image segmentation model to obtain the segmentation result. This invention addresses the problem of insufficient accuracy in medical image segmentation models trained in the prior art. It improves the segmentation accuracy of medical image segmentation models by training them through generative adversarial methods.

[0052] Figure 4 This is a schematic diagram of a medical image segmentation device provided in an embodiment of the present invention. The embodiment of the present invention is applicable to scenarios where medical images are segmented. The device can be implemented by software and / or hardware and integrated into a computer device with application development capabilities.

[0053] like Figure 4As shown, the medical image segmentation device includes: a medical image acquisition module 310 and a medical image segmentation module 320.

[0054] The medical image acquisition module 310 is used to acquire the medical image to be segmented; the medical image segmentation module 320 is used to input the medical image to be segmented into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented.

[0055] The technical solution provided by this invention obtains the segmentation result of the medical image to be segmented by inputting the medical image to be segmented into a target medical image segmentation model. The target medical image segmentation model is a model trained through a generative adversarial approach based on preset source domain medical sample images and preset target domain medical sample images. The preset target domain medical sample images are medical sample images that have the same region to be segmented as the preset source domain medical images but possess at least one different image feature. This invention solves the problem of insufficient accuracy in medical image segmentation models trained by existing technologies. It improves the segmentation accuracy of medical image segmentation models by training them through a generative adversarial approach.

[0056] In one optional embodiment, the medical image segmentation device further includes: a medical image segmentation model training module, configured to: take a preset source domain medical sample image with sample labels and a preset target domain medical sample image without labels as a set of sample image pairs; input the sample image pairs into an initial medical image segmentation model, and extract features from the sample image pairs through the encoder of the initial medical image segmentation model to obtain source domain image features and target domain image features; input the source domain image features and target domain image features into a preset image domain recognizer and a decoder of the initial medical image segmentation model to obtain corresponding image domain recognition results and image segmentation results; adjust the parameters of the initial medical image segmentation model based on the image domain recognition results and image segmentation results to obtain a target medical image segmentation model.

[0057] In one optional implementation, the medical image segmentation model training module is specifically used to: determine an image domain recognition loss function based on the image domain recognition results of a preset target domain medical sample image and a preset source domain medical sample image; determine a target domain image segmentation loss function based on the image segmentation results of the preset target domain medical sample image; determine a source domain image segmentation loss function based on the image segmentation results of the preset source domain medical sample image; and adjust the parameter values ​​of the encoder and decoder based on the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function to obtain the target medical image segmentation model.

[0058] In one optional implementation, the medical image segmentation model training module is specifically used to: sum the inverses of the target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function to obtain the encoder loss function; determine the encoder parameter adjustment value based on the encoder loss function; and adjust the encoder parameter value based on the encoder parameter adjustment value.

[0059] In one optional implementation, the medical image segmentation model training module is specifically used to: sum the target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function to obtain the decoder loss function; determine the decoder parameter adjustment value based on the decoder loss function; and adjust the decoder parameter value based on the decoder parameter adjustment value.

[0060] In one optional implementation, the medical image segmentation model training module is specifically used to: calculate the image pixel entropy of each pixel in the preset target domain medical sample image based on the image segmentation result of the preset target domain medical sample image; and determine the target domain image segmentation loss function based on the image pixel entropy.

[0061] In one alternative implementation, the target medical image segmentation model includes a U-Net network structure.

[0062] The medical image segmentation apparatus provided in this embodiment of the invention can execute the medical image segmentation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0063] Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Figure 5 A block diagram of an exemplary computer device 12 suitable for implementing embodiments of the present invention is shown. Figure 5 The computer device 12 shown is merely an example and should not be construed as limiting the functionality or scope of the embodiments of the present invention. The computer device 12 can be any terminal device with computing capabilities and can be integrated into a medical image segmentation device.

[0064] like Figure 5 As shown, the computer device 12 is represented in the form of a general-purpose computing device. The components of the computer device 12 may include, but are not limited to: one or more processors or processing units 16, system memory 28, and a bus 18 connecting different system components (including system memory 28 and processing unit 16).

[0065] Bus 18 can be one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0066] Computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by computer device 12, including volatile and non-volatile media, removable and non-removable media.

[0067] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache 32. Computer device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 5 Not shown; usually referred to as a "hard drive"). Although Figure 5 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0068] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 42 typically perform the functions and / or methods described in the embodiments of the present invention.

[0069] Computer device 12 can also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), and with one or more devices that enable a user to interact with the computer device 12, and / or with any device that enables the computer device 12 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed via input / output (I / O) interface 22. Furthermore, computer device 12 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 20. As shown, network adapter 20 communicates with other modules of computer device 12 via bus 18. It should be understood that, although... Figure 5 As not shown, it can be used in conjunction with computer device 12 with other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0070] Processing unit 16 executes various functional applications and data processing by running programs stored in system memory 28, such as implementing the medical image segmentation method provided in this embodiment, which includes:

[0071] Acquire the medical image to be segmented;

[0072] The medical image to be segmented is input into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented;

[0073] The target medical image segmentation model is a model trained by generating adversarial methods based on a preset source domain medical sample image and a preset target domain medical sample image. The preset target domain medical sample image is a medical sample image that has the same region to be segmented as the preset source domain medical image and has at least one different image feature.

[0074] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the medical image segmentation method as provided in any embodiment of the present invention, including:

[0075] Acquire the medical image to be segmented;

[0076] The medical image to be segmented is input into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented;

[0077] The target medical image segmentation model is a model trained by generating adversarial methods based on a preset source domain medical sample image and a preset target domain medical sample image. The preset target domain medical sample image is a medical sample image that has the same region to be segmented as the preset source domain medical image and has at least one different image feature.

[0078] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. For example, a computer-readable storage medium can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0079] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0080] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0081] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0082] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computing device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0083] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A medical image segmentation method, characterized in that, include: Acquire the medical image to be segmented; The medical image to be segmented is input into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented; The target medical image segmentation model is a model trained by generating adversarial methods based on a preset source domain medical sample image and a preset target domain medical sample image. The preset target domain medical sample image is a medical sample image that has the same region to be segmented as the preset source domain medical image and has at least one different image feature. The process of training the target medical image segmentation model based on preset source domain medical sample images and preset target domain medical sample images through generative adversarial methods includes: A set of sample image pairs is created by combining a pre-defined source domain medical sample image with sample labels and a pre-defined target domain medical sample image without labels. The sample image pairs are respectively input into the initial medical image segmentation model, and the encoder of the initial medical image segmentation model is used to extract features from the sample image pairs to obtain source domain image features and target domain image features. The source domain image features and the target domain image features are input into the decoder of the preset image domain recognizer and the initial medical image segmentation model to obtain the corresponding image domain recognition results and image segmentation results, respectively; wherein, the preset image domain recognizer is a preset recognizer used to recognize the preset source domain medical sample image and the preset target domain medical sample image; The parameters of the initial medical image segmentation model are adjusted based on the image domain recognition results and the image segmentation results to obtain the target medical image segmentation model; The process of adjusting the parameters of the initial medical image segmentation model based on the image domain recognition result and the image segmentation result to obtain the target medical image segmentation model includes: The image domain recognition loss function is determined based on the image domain recognition results of the preset target domain medical sample image and the image domain recognition results of the preset source domain medical sample image. The target domain image segmentation loss function is determined based on the image segmentation results of the medical sample image in the preset target domain; the target domain image segmentation loss function is determined based on the image pixel entropy of each pixel in the medical sample image in the preset target domain calculated from the image segmentation results. The source domain image segmentation loss function is determined based on the image segmentation results of the preset source domain medical sample images; the source domain image segmentation loss function is calculated based on the image segmentation results of the preset source domain medical sample images and the sample labels of the preset source domain medical sample images. The parameter values ​​of the encoder and the decoder are adjusted based on the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function to obtain the target medical image segmentation model; The parameter values ​​of the encoder are adjusted based on the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function, including: The encoder loss function is obtained by summing the inverses of the target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function. The encoder parameter adjustment value is determined based on the encoder loss function, and the encoder parameter values ​​are adjusted based on the encoder parameter adjustment value. The parameter values ​​of the decoder are adjusted based on the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function, including: The target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function are summed to obtain the decoder loss function; The decoder parameter adjustment value is determined based on the decoder loss function, and the decoder parameter value is adjusted based on the decoder parameter adjustment value.

2. The method according to claim 1, characterized in that, The step of determining the target domain image segmentation loss function based on the image segmentation results of the preset target domain medical sample image includes: Calculate the image pixel entropy of each pixel in the preset target domain medical sample image based on the image segmentation results of the preset target domain medical sample image; The target domain image segmentation loss function is determined based on the image pixel entropy.

3. The method according to claim 1, characterized in that, The target medical image segmentation model includes a U-Net network structure.

4. A medical image segmentation device, characterized in that, The device includes: The medical image acquisition module is used to acquire the medical image to be segmented. The medical image segmentation module is used to input the medical image to be segmented into the target medical image segmentation model to obtain the segmentation result of the medical image to be segmented; The target medical image segmentation model is a model trained by generating adversarial methods based on a preset source domain medical sample image and a preset target domain medical sample image. The preset target domain medical sample image is a medical sample image that has the same region to be segmented as the preset source domain medical image and has at least one different image feature. The process of training the target medical image segmentation model based on preset source domain medical sample images and preset target domain medical sample images through generative adversarial methods includes: A set of sample image pairs is created by combining a pre-defined source domain medical sample image with sample labels and a pre-defined target domain medical sample image without labels. The sample image pairs are respectively input into the initial medical image segmentation model, and the encoder of the initial medical image segmentation model is used to extract features from the sample image pairs to obtain source domain image features and target domain image features. The source domain image features and the target domain image features are input into the decoder of the preset image domain recognizer and the initial medical image segmentation model to obtain the corresponding image domain recognition results and image segmentation results, respectively; wherein, the preset image domain recognizer is a preset recognizer used to recognize the preset source domain medical sample image and the preset target domain medical sample image; The parameters of the initial medical image segmentation model are adjusted based on the image domain recognition results and the image segmentation results to obtain the target medical image segmentation model; The process of adjusting the parameters of the initial medical image segmentation model based on the image domain recognition result and the image segmentation result to obtain the target medical image segmentation model includes: The image domain recognition loss function is determined based on the image domain recognition results of the preset target domain medical sample image and the image domain recognition results of the preset source domain medical sample image. The target domain image segmentation loss function is determined based on the image segmentation results of the medical sample image in the preset target domain; the target domain image segmentation loss function is determined based on the image pixel entropy of each pixel in the medical sample image in the preset target domain calculated from the image segmentation results. The source domain image segmentation loss function is determined based on the image segmentation results of the preset source domain medical sample images; the source domain image segmentation loss function is calculated based on the image segmentation results of the preset source domain medical sample images and the sample labels of the preset source domain medical sample images. The parameter values ​​of the encoder and the decoder are adjusted based on the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function to obtain the target medical image segmentation model; The parameter values ​​of the encoder are adjusted based on the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function, including: The encoder loss function is obtained by summing the inverses of the target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function. The encoder parameter adjustment value is determined based on the encoder loss function, and the encoder parameter values ​​are adjusted based on the encoder parameter adjustment value. The parameter values ​​of the decoder are adjusted based on the image domain recognition loss function, the target domain image segmentation loss function, and the source domain image segmentation loss function, including: The target domain image segmentation loss function, the source domain image segmentation loss function, and the image domain recognition loss function are summed to obtain the decoder loss function; The decoder parameter adjustment value is determined based on the decoder loss function, and the decoder parameter value is adjusted based on the decoder parameter adjustment value.

5. A computer device, characterized in that, The computer device includes: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the medical image segmentation method as described in any one of claims 1-3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the medical image segmentation method as described in any one of claims 1-3.

Images