Image segmentation method and device, medium and product
By using pre-trained parameter models and training samples with added perturbation information in the medical image segmentation model, the problem of insufficient anti-interference ability of the region of interest segmentation model in the prior art is solved, and accurate segmentation and goodness of noisy medical images are achieved. Generalization ability.
Patent Information
- Application Number
- CN202411202254.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-12-17
AI Technical Summary
Existing segmentation models of areas of interest are difficult to complete accurate segmentation in medical images containing perturbation information because the model is more concerned about whether the model has learned enough segmentation experience while not paying enough attention to anti-interference ability.
By using pre-trained segmentation model, the initial parameters at the beginning of model training are determined based on the pre-trained parameter model, and are trained in combination with reference images with added perturbation information, ensuring that the model has good anti-interference ability and image recovery ability during the training process, so as to focus on Learn from segmentation experience in the area of interest.
It realizes accurate segmentation of the region of interest on image blocks containing various interference information, improves the generalization ability of the segmentation model, and ensures accurate target segmentation results on the medical images to be segmented.
Smart Images
Figure CN119151958A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medical image processing, and in particular to an image segmentation method, device, medium and product. Background Art
[0002] Existing region of interest segmentation models usually find it difficult to complete the region of interest segmentation of medical images containing perturbation information. The reason is that during the model training process, more attention is paid to whether the model has learned sufficient region of interest segmentation experience, but little attention is paid to whether it has learned anti-interference experience.
[0003] Therefore, it is necessary to provide an image segmentation method for noisy medical images based on a segmentation model. Summary of the invention
[0004] The present invention provides an image segmentation method, device, medium and product to solve the problem that the existing image segmentation method cannot determine the image segmentation result of a medical image containing disturbance information based on a segmentation model.
[0005] According to one aspect of the present invention, there is provided an image segmentation method, comprising:
[0006] Acquire a medical image to be segmented;
[0007] Using a pre-trained segmentation model to perform region of interest segmentation on the medical image to be segmented, to obtain a target segmentation result of the medical image to be segmented;
[0008] Among them, the initial model parameters of the segmentation model at the beginning of model training are determined according to the model parameters of the pre-trained parameter model, the first training sample used to train the parameter model includes a reference image and a reference image with disturbance information added, and the parameter model has the same network structure as the segmentation model.
[0009] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:
[0010] at least one processor; and
[0011] a memory communicatively connected to the at least one processor; wherein,
[0012] The memory stores a computer program executable by the at least one processor. The computer program is executed by the at least one processor so that the at least one processor can perform the image segmentation method described in any embodiment.
[0013] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the image segmentation method described in any embodiment of the present invention when executed.
[0014] According to another aspect of the present invention, a computer program product is provided. The computer program product comprises a computer program. When the computer program is executed by a processor, the image segmentation method according to any embodiment is implemented.
[0015] The technical solution of the embodiment of the present invention determines the initial model parameters of the segmentation model at the beginning of model training according to the model parameters of the pre-trained parameter model, so that the segmentation model has good anti-interference ability and image restoration ability at the beginning of model training, and makes the segmentation model more focused on learning the experience of region of interest segmentation during the training process, so that after the training is completed, it can complete accurate region of interest segmentation for image blocks containing various interference information, and ensure that the pre-trained segmentation model has good generalization ability. Therefore, by using the pre-trained segmentation model to segment the medical image to be segmented, accurate target segmentation results can be obtained.
[0016] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0018] Figure 1 is a flow chart of an image segmentation method provided according to an embodiment of the present invention;
[0019] Figure 2 is a flowchart of a model training method provided according to an embodiment of the present invention;
[0020] Figure 3 is a structural schematic diagram of an image segmentation device provided according to an embodiment of the present invention;
[0021] Figure 4 It is a schematic diagram of the structure of an electronic device for implementing the image segmentation method according to an embodiment of the present invention. DETAILED DESCRIPTION
[0022] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0024] Figure 1 This is a flow chart of an image segmentation method provided by an embodiment of the present invention. This embodiment is applicable to the case where segmentation model training is completed with fewer medical images. The method can be executed by a model training device, which can be implemented in the form of hardware and / or software. The model training device can be configured in a processor of an electronic device. Figure 1 As shown, the method includes:
[0025] S110: Obtain a medical image to be segmented.
[0026] When a medical image is detected, first determine whether the image format is correct, such as whether it is a three-dimensional image, whether it is in nii format, etc.; if not, output a prompt message. If yes, cut the medical image to be segmented into at least two image blocks. By cutting the medical image to be segmented into at least two image blocks and segmenting each image block one by one, the single data processing amount of the pre-trained segmentation model can be reduced.
[0027] S120. Use a pre-trained segmentation model to perform region of interest segmentation on the medical image to be segmented, and obtain a target segmentation result of the medical image to be segmented, wherein initial model parameters of the segmentation model at the beginning of model training are determined according to model parameters of the pre-trained parameter model, and the first training sample used to train the parameter model includes a reference image and a reference image with disturbance information added, the parameter model has the same network structure as the segmentation model, and the reference image is a part of the medical image.
[0028] In one embodiment, after obtaining at least two image blocks, the at least two image blocks are resampled respectively to update the at least two image blocks, and the voxel size of the updated image blocks is different from the voxel size of the image blocks. This embodiment changes the voxel size of the image blocks by resampling, thereby changing the resolution of the image blocks. Preferably, each image block is resampled to obtain a corresponding target image block, and the voxel size of the target image block is smaller than the voxel size of each image block.
[0029] In one embodiment, before resampling, it is determined whether the image block includes a blank area, and if so, the blank area is deleted to update the image block. Alternatively, after resampling, it is determined whether the updated image block includes a blank area, and if so, the blank area is deleted to update the image block.
[0030] The at least two image blocks are sequentially input into the pre-trained segmentation model so that the pre-trained segmentation model performs region of interest segmentation on each image block to obtain region of interest segmentation results of each image block. In the region of interest segmentation results, the voxel value of the voxel corresponding to the region of interest is 1, and the voxel value of the voxel corresponding to the non-region of interest is 0. The region of interest is the area of user attention, such as the lesion area.
[0031] Among them, the initial model parameters of the segmentation model are determined according to the model parameters of the pre-trained parameter model. In one embodiment, since the network structure of the parameter model is the same as that of the segmentation model, the model parameter data of the pre-trained parameter model is assigned to the model parameters of the segmentation model, so that the segmentation model has better anti-interference ability, image restoration ability and feature extraction ability at the beginning of model training, so that it can better eliminate various interference information in the training stage, quickly learn various regions of interest segmentation experience, effectively improve the stability of segmentation model training, and improve the feature extraction ability of the segmentation model.
[0032] The first training sample includes a reference image and a reference image with disturbance information added. During the training process, the parameter model uses the reference image as a reference and performs image restoration processing on the reference image with disturbance information added until the image restoration result meets the set evaluation condition.
[0033] The reference image is a part of the complete medical image, and the size of the reference image may be the same as or different from the size of each image block in the at least two image blocks.
[0034] In one embodiment, after the segmentation results of the regions of interest corresponding to each image block are determined, the segmentation results of the regions of interest of each image block are resampled to update the segmentation results of the regions of interest of each image block, and the voxel size of the updated segmentation results of the regions of interest is the same as the voxel size of the medical image to be segmented. This embodiment restores the voxel size of the segmentation results of the regions of interest to the voxel size of the medical image to be segmented through resampling, so as to facilitate the subsequent determination of the target segmentation result and the use of the target segmentation results of the regions of interest, such as registration with other positioning images.
[0035] In one embodiment, the positional relationship between the region of interest segmentation results corresponding to each image block is determined based on the relative positional relationship between the image blocks, and the region of interest segmentation results corresponding to each image block are spliced together based on the positional relationship between the region of interest segmentation results to obtain the target segmentation result.
[0036] In one embodiment, a base plate image corresponding to the medical image is determined, and the voxel value of each voxel on the base plate image is zero; according to the position of the image block corresponding to each of the region of interest segmentation results on the medical image to be segmented, each of the region of interest segmentation results is added to the corresponding position of the base plate image to obtain a target segmentation result. This embodiment can quickly and accurately determine the target region of interest segmentation result.
[0037] In one embodiment, the reference model and the segmentation model adopt the same model structure, including a first dual convolution layer, a second dual convolution layer, a first downsampling module, a second downsampling module, a third downsampling module, a fourth downsampling module, a first upsampling module, a second upsampling module, a third upsampling module and a fourth upsampling module connected in sequence, and the output end of the fourth downsampling module is also connected to the input end of the first upsampling module, the output end of the third downsampling module is also connected to the input end of the second upsampling module, the output end of the second downsampling module is also connected to the input end of the third upsampling module, and the output end of the first downsampling module is also connected to the output end of the fourth upsampling module. Among them, the dual convolution layer includes a three-dimensional convolution layer, a normalization layer, an activation function layer, a three-dimensional convolution layer, a normalization layer, and an activation function layer connected in sequence. The downsampling module includes a dual convolution layer and a maximum pooling downsampling layer connected in sequence. The upsampling module includes a transposed convolution layer and a dual convolution layer connected in sequence.
[0038] The technical solution provided by the embodiment of the present invention determines the initial model parameters of the segmentation model at the beginning of model training according to the model parameters of the pre-trained parameter model, so that the segmentation model has good anti-interference ability and image restoration ability at the beginning of model training, and makes the segmentation model more focused on learning the experience of region of interest segmentation during the training process, so that after the training is completed, it can complete accurate region of interest segmentation for image blocks containing various interference information, and ensure that the pre-trained segmentation model has good generalization ability. Therefore, by using the pre-trained segmentation model to segment the medical image to be segmented, accurate target segmentation results can be obtained.
[0039] Figure 2 This is a flow chart of a model training method provided by an embodiment of the present invention, and this embodiment is used to illustrate the training method of the parameter model. Figure 1 As shown, the method includes:
[0040] S210 , obtaining a first training sample set including a plurality of first training samples and a second training sample set including a plurality of second training samples, wherein the second training samples include a target image and a segmentation result of a region of interest of the target image.
[0041] In one embodiment, the reference image is a complete medical image of the target object.
[0042] In one embodiment, the reference image is a part of a complete medical image of the target object. This embodiment can reduce the number of medical images required when determining the first training sample set and the second training sample set.
[0043] It is understandable that, when the image size is the same, the smaller the voxel size, the greater the number of voxels, the larger the image data, and the higher the accuracy of the corresponding region of interest segmentation result. In actual use, it can be determined whether the medical image needs to be resampled to convert its voxel size to the required voxel size based on the voxel size of the medical image output by the imaging device and the accuracy of the required region of interest segmentation result.
[0044] Optionally, a medical image of the target object is obtained and the medical image is divided into at least two image blocks; the image blocks are resampled to obtain a reference image so that the voxel size of the reference image is smaller than the voxel size of the image blocks. This embodiment reduces the voxel size by resampling and improves the resolution of the image blocks, and the improvement of the image block resolution helps to improve the model training effect.
[0045] In one embodiment, after the image block is determined, a random spatial transformation is performed on the image block to update the image block, such as random rotation, random cropping, random deformation, random mirroring, etc. The random spatial transformation increases the diversity of the first training samples included in the first training sample set, increases the complexity of the model input samples, and alleviates the overfitting of the model.
[0046] The medical image is a clinical diagnosis medical image or a surgical path planning image, such as a CT image, an MRI image or a PET image, especially a PET image with fewer image samples. In one embodiment, the reference image and the target image are both image blocks of the PET image.
[0047] In one embodiment, disturbance information is added to the reference image based on a random noise generation method to obtain a reference image with disturbance information added. The disturbance information includes but is not limited to at least one of random noise, random blur, random brightness, and random contrast.
[0048] The region of interest segmentation result of the target image may be selected as a manually determined region of interest segmentation result.
[0049] In one embodiment, the reference image has the same size as the target image; the voxel size of the reference image has the same voxel size as the target image. This embodiment can improve the model parameter data migration effect, that is, the initial segmentation model can better complete denoising and region of interest segmentation during the region of interest segmentation process.
[0050] S220: Complete the training of the parameter model based on the first training sample set to obtain a pre-trained parameter model.
[0051] Each first training sample in the first training sample set is input into the parameter model, and the parameter model performs denoising on the reference image with disturbance information added to the first training sample to obtain a denoising result, and then adjusts the model parameters according to the comparison result between the denoising result and the reference image, and learns image restoration experience until the difference between the image restoration result and the reference image meets the set constraint conditions, thereby obtaining a pre-trained parameter model. It can be understood that the pre-trained parameter model has good image restoration capabilities.
[0052] S230, initializing the model parameters of the segmentation model based on the model parameters of the pre-trained parameter model to obtain an initial segmentation model, wherein the segmentation model has the same network structure as the parameter model.
[0053] Since the network structure of the segmentation model is the same as that of the parameter model, the model parameter data of the parameter model is assigned to the model parameters of the segmentation model to complete the model parameter initialization of the segmentation model and obtain the initial segmentation model.
[0054] It can be understood that since the segmentation model and the parameter model have the same network structure, after the model parameter data of the parameter model is assigned to the segmentation model to obtain the initial segmentation model, the initial segmentation model has the same image restoration ability as the parameter model.
[0055] S240: Complete the training of the initial segmentation model based on the second training sample set to obtain a pre-trained segmentation model.
[0056] The initial segmentation model is trained using each second training sample in the second training sample set. During the training process, the initial segmentation model performs region of interest segmentation on the target image in the second training sample to obtain a model region of interest segmentation result; the model region of interest segmentation result is compared with the region of interest segmentation result in the second training sample, and the model parameters are adjusted according to the comparison result until the comparison result meets the set conditions.
[0057] It is understandable that the pre-trained segmentation model has good ROI segmentation and image restoration capabilities. Therefore, regardless of whether the medical image contains disturbance information or what kind of disturbance information it contains, the pre-trained segmentation model can accurately segment the ROI and obtain accurate ROI segmentation results.
[0058] The technical solution provided by the embodiment of the present invention first completes the training of the parameter model based on the first training sample set to obtain a pre-trained parameter model, then assigns the model parameter data of the pre-trained parameter model to the segmentation model, completes the model parameter initialization of the segmentation model, and obtains the initial segmentation model, so that the initial segmentation model has good image restoration capabilities at the beginning of model training; finally, the model training of the initial segmentation model is completed based on the second training sample set to obtain the pre-trained segmentation model. The training process enables the pre-trained segmentation model to have good image restoration capabilities while also learning good region of interest segmentation experience. Therefore, regardless of whether the medical image contains disturbance information or what kind of disturbance information it contains, the pre-trained segmentation model can accurately segment the region of interest and obtain accurate region of interest segmentation results.
[0059] Figure 3 Schematic diagram of the structure of the image segmentation device provided by the embodiment of the present invention. Figure 3 As shown, the device comprises:
[0060] A segmentation module 31, used for acquiring a medical image to be segmented;
[0061] A segmentation module 32 is used to segment the region of interest of the medical image to be segmented using a pre-trained segmentation model to obtain a target segmentation result of the medical image to be segmented;
[0062] Among them, the initial model parameters of the segmentation model at the beginning of model training are determined according to the model parameters of the pre-trained parameter model, the first training sample used to train the parameter model includes a reference image and a reference image with disturbance information added, and the parameter model has the same network structure as the segmentation model.
[0063] In one embodiment, the segmentation module is specifically used to:
[0064] Dividing the medical image to be segmented into at least two image blocks;
[0065] Using a pre-trained segmentation model to perform region of interest segmentation on each of the at least two image blocks to obtain a region of interest segmentation result for each of the image blocks,
[0066] According to the segmentation results of the regions of interest corresponding to the image blocks, a target segmentation result of the medical image to be segmented is determined.
[0067] In one embodiment, the device further comprises a model training module, wherein the model training module is used to:
[0068] Acquire a first training sample set including a plurality of first training samples and a second training sample set including a plurality of second training samples, wherein the second training samples include a target image and a segmentation result of a region of interest of the target image;
[0069] Complete the training of the parameter model based on the first training sample set to obtain a pre-trained parameter model;
[0070] Initialize the model parameters of the segmentation model based on the model parameters of the pre-trained parameter model to obtain an initial segmentation model;
[0071] The training of the initial segmentation model is completed based on the second training sample set to obtain a pre-trained segmentation model.
[0072] In one embodiment, the reference image has the same size as the target image;
[0073] The reference image has the same size as the target image and both are part of the medical image.
[0074] In one embodiment, the model training module is further used to:
[0075] Acquire a medical image of a target object, and divide the medical image into at least two image blocks;
[0076] The image block is resampled to obtain a reference image, so that a voxel size of the reference image is smaller than a voxel size of the image block.
[0077] In one embodiment, both the reference image and the target image are PET images.
[0078] In one embodiment, the segmentation module 31 is further used for:
[0079] Resampling the at least two image blocks respectively to update the at least two image blocks, wherein the voxel size of the updated image blocks is different from the voxel size of the image blocks;
[0080] In one embodiment, the result module 33 is further used to:
[0081] The region of interest segmentation result of each image block is resampled to update the region of interest segmentation result of each image block, and the voxel size of the updated region of interest segmentation result is the same as the voxel size of the medical image to be segmented.
[0082] In one embodiment, the result module is specifically used to:
[0083] Determine a base plate image corresponding to the medical image, wherein the voxel value of each voxel on the base plate image is zero;
[0084] According to the position of the image block corresponding to each of the region of interest segmentation results on the medical image to be segmented, each of the region of interest segmentation results is added to the corresponding position of the base plate image to obtain a target segmentation result.
[0085] The technical solution provided by the embodiment of the present invention determines the initial model parameters of the segmentation model at the beginning of model training according to the model parameters of the pre-trained parameter model, so that the segmentation model has good anti-interference ability and image restoration ability at the beginning of model training, and makes the segmentation model more focused on learning the experience of region of interest segmentation during the training process, so that after the training is completed, it can complete accurate region of interest segmentation for image blocks containing various noises, and ensure that the pre-trained segmentation model has good generalization ability. Therefore, by using the pre-trained segmentation model to segment the medical image to be segmented, accurate target segmentation results can be obtained.
[0086] The image segmentation device provided in the embodiment of the present invention can execute the image segmentation method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0087] Figure 4A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.
[0088] like Figure 4 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.
[0089] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.
[0090] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the image segmentation method.
[0091] In some embodiments, the image segmentation method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the image segmentation method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the image segmentation method in any other appropriate manner (e.g., by means of firmware).
[0092] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0093] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.
[0094] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0095] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).
[0096] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.
[0097] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.
[0098] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.
[0099] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. An image segmentation method, characterized in that: include: Acquire a medical image to be segmented; Using a pre-trained segmentation model to perform region of interest segmentation on the medical image to be segmented, to obtain a target segmentation result of the medical image to be segmented; Among them, the initial model parameters of the segmentation model at the beginning of model training are determined according to the model parameters of the pre-trained parameter model, the first training sample used to train the parameter model includes a reference image and a reference image with disturbance information added, and the parameter model has the same network structure as the segmentation model.
2. The method according to claim 1, characterized in that The using the pre-trained segmentation model to perform region of interest segmentation on the medical image to be segmented to obtain a region of interest segmentation result includes: Dividing the medical image to be segmented into at least two image blocks; Using a pre-trained segmentation model to perform region of interest segmentation on each of the at least two image blocks to obtain a region of interest segmentation result for each of the image blocks, According to the segmentation results of the regions of interest corresponding to the image blocks, a target segmentation result of the medical image to be segmented is determined.
3. The method according to claim 1, characterized in that Determine the pre-trained segmentation model by the following method; Acquire a first training sample set including a plurality of first training samples and a second training sample set including a plurality of second training samples, wherein the second training samples include a target image and a segmentation result of a region of interest of the target image; Complete the training of the parameter model based on the first training sample set to obtain a pre-trained parameter model; Initialize the model parameters of the segmentation model based on the model parameters of the pre-trained parameter model to obtain an initial segmentation model; The training of the initial segmentation model is completed based on the second training sample set to obtain a pre-trained segmentation model.
4. The method according to claim 3, characterized in that The reference image has the same size as the target image and both are part of the medical image; The voxel size of the reference image is the same as the voxel size of the target image.
5. The method according to claim 3, characterized in that: The reference image and the target image are both PET images.
6. The method according to claim 2, characterized in that After obtaining the medical image to be segmented, the method further includes: Resampling the at least two image blocks respectively to update the at least two image blocks, wherein the voxel size of the updated image blocks is different from the voxel size of the image blocks; After obtaining the segmentation result of the region of interest of each image block, the method further includes: The region of interest segmentation result of each image block is resampled to update the region of interest segmentation result of each image block, and the voxel size of the updated region of interest segmentation result is the same as the voxel size of the medical image to be segmented.
7. The method according to claim 2, characterized in that: Determining the target segmentation result of the medical image to be segmented according to the segmentation results of the regions of interest corresponding to the image blocks includes: Determine a base plate image corresponding to the medical image, wherein the voxel value of each voxel on the base plate image is zero; According to the position of the image block corresponding to each of the region of interest segmentation results on the medical image to be segmented, each of the region of interest segmentation results is added to the corresponding position of the base plate image to obtain a target segmentation result.
8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the image segmentation method according to any one of claims 1 to 7.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the image segmentation method according to any one of claims 1 to 7 when executed.
10. A computer program product, characterized in that The computer program product comprises a computer program, which, when executed by a processor, implements the image segmentation method according to any one of claims 1 to 7.