Slice image processing method and apparatus, and network device

Abstract

The embodiment of the application provides a kind of slice image processing method, device and network equipment, the method comprises: obtaining scanning data, scanning data includes the information corresponding to the multiple layers of slice of scanning object;The initialization of scanning data is obtained with the multiple initialization images corresponding to multiple layers of slice;According to multiple initialization images, iterative processing is carried out, and multiple slice images are obtained, each time iterative processing includes analyzing the noise data of input image, and the output image is obtained after processing input image according to noise data, the noise data meets self-consistency, and the self-consistency of noise data is that the data obtained by re-splitting after mixing the noise data corresponding to multiple initialization images meets consistency between noise data;The present scheme uses multi-layer excitation technology to reduce the time of nuclear magnetic resonance acquisition data, improves sampling efficiency, and the method used can improve reconstruction quality.

Description

Technical Field

[0001] This application relates to the field of computer information technology, and more specifically, to a method, apparatus, and network device for processing sliced ​​images. Background Technology

[0002] Magnetic resonance imaging (MRI) is the process of acquiring slice scan data from various locations of a target site (such as the heart) and then reconstructing them into slice images.

[0003] Existing solutions typically require scanning each slice to obtain slice images, resulting in slow data acquisition. Summary of the Invention

[0004] This application provides a method, apparatus, and network device for processing sliced ​​images, which can generate multiple slice data from a single scan, thereby reducing the number of scans and improving data acquisition efficiency. The technical solutions are as follows:

[0005] In a first aspect, this application provides a slice image processing method, the method comprising: acquiring scan data, the scan data including information corresponding to multiple slices of the scanned object; initializing the scan data to obtain multiple initialized images corresponding to the multiple slices; performing iterative processing based on the multiple initialized images to obtain multiple slice images, each iteration including analyzing noise data of the input image, and processing the input image based on the noise data to obtain an output image, wherein the noise data conforms to self-consistency, and the self-consistency of the noise data is that the data obtained by mixing and re-splitting the noise data corresponding to the multiple initialized images conforms to the noise data.

[0006] Furthermore, the iterative processing based on multiple initial images to obtain multiple slice images includes: determining the time step and multiple input images for the current iteration, wherein the multiple input images are multiple output images or multiple initial images obtained from the previous iteration; inputting the multiple input images and the iteration time step into a pre-trained processing model to obtain self-consistent noise data, wherein the processing model is used to analyze the noise contained in the images based on the input multiple images; completing the current iteration based on the multiple input images and noise data to determine multiple output images; and performing the next round of iteration based on the multiple output images until multiple slice images are obtained.

[0007] Furthermore, the step of completing the iterative processing based on multiple input images and noise data to determine multiple output images includes: generating randomized noise and performing self-consistency processing on the randomized noise to obtain randomized noise data; and completing the iterative processing based on multiple input images, noise data, and randomized noise data to determine multiple output images.

[0008] Furthermore, the step of completing the iterative processing based on multiple input images, noise data, and random noise data to determine multiple output images includes: determining the round data corresponding to the input images and analyzing the difference data between the round data and the scan data; and completing the iterative processing based on multiple input images, noise data, random noise data, and difference data to determine multiple output images.

[0009] Furthermore, the multiple input images conform to self-consistency. Self-consistency of multiple input images means that the image obtained after mixing and re-splitting the multiple input images conforms to the consistency of the multiple input images. The step of completing the current iteration processing based on multiple input images, noise data, random noise data, and difference data to determine multiple output images includes: determining the fusion ratio corresponding to the multiple input images to determine the fusion data corresponding to the multiple input images; and completing the current iteration processing based on multiple input images, noise data, random noise data, difference data, and fusion data to determine multiple output images.

[0010] Furthermore, the method also includes the step of training the processing model: acquiring multiple training images and iteratively training the multiple training images to determine multiple training noise images; wherein, each iteration of training includes analyzing the training noise data of the training input image, and processing the training input image according to the noise data to obtain the training output image; the training noise data conforms to self-consistency, and the self-consistency of the training noise data is that the data obtained by mixing and re-splitting the training noise data corresponding to multiple training input images is consistent with the training noise data.

[0011] Furthermore, the iterative training of multiple training images includes: determining multiple training input images for the current iteration, wherein the multiple training input images are the training output data of the previous iteration or multiple training images; generating training noise data corresponding to the multiple training input images for the current iteration; fusing the multiple training input images with the training noise of the current iteration for the next iteration, until multiple training noise images are determined; and training the processing model based on the multiple training noise images, the multiple training images, and the training noise data from each iteration, until a trained processing model is obtained.

[0012] Secondly, this application provides a slice image processing apparatus, the apparatus comprising: a scan data acquisition module for acquiring scan data, the scan data including information corresponding to multiple slices of the scanned object; a scan data initialization module for initializing the scan data to obtain multiple initial images corresponding to the multiple slices; and a slice image generation module for iterative processing based on the multiple initial images to obtain multiple slice images, each iteration including analyzing noise data of the input image and processing the input image based on the noise data to obtain an output image, wherein the noise data conforms to self-consistency, and the self-consistency of the noise data is that the data obtained by mixing and re-splitting the noise data corresponding to the multiple initial images conforms to the noise data.

[0013] Thirdly, this application provides a network device, including: a memory, a transceiver, and a processor; wherein the memory is used to store a computer program; the transceiver is used to send and receive data under the control of the processor; and the processor is used to read the computer program in the memory and execute the method as described in the first aspect.

[0014] Fourthly, this application provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0015] The beneficial effects of the technical solution provided in this application are:

[0016] The proposed solution can be applied to nuclear magnetic resonance (NMR) image acquisition scenarios, reducing the number of NMR data acquisitions and improving sampling efficiency. Specifically, existing solutions, to avoid noise generated by simultaneously acquiring data from various slices, typically use NMR technology to scan each slice of the object separately, then acquire data from each slice for reconstruction. This solution can acquire scan data containing information about overlapping and ambiguity of multiple slices of the object. After acquiring the scan data, it can be initialized into an initial image corresponding to the multiple slices of the object, and iterative processing is used to maintain image consistency during noise removal, thus obtaining a multi-slice image corresponding to the multiple slices of the object, facilitating subsequent 3D reconstruction and other processing. This solution can acquire scan data of multiple slices of the object simultaneously using multi-slice scanning for separation, denoising, and other processing, reducing the number of samplings and improving sampling efficiency. This solution utilizes multi-slice excitation technology to reduce NMR data acquisition time and improve sampling efficiency; the method used can improve reconstruction quality. Specifically, this solution can pre-train a processing model, which analyzes the noise contained in the input multiple images for subsequent denoising processing. This scheme can acquire scan data, which includes information corresponding to multiple slices of the scanned object; initialize the scan data to obtain multiple initialized images corresponding to the multiple slices; then, perform multiple iterative processing based on the multiple initialized images and a pre-trained processing model to obtain multiple slice images. Each iteration includes analyzing the noise data of the input image based on the pre-trained processing model, and processing the input image based on the noise data to obtain the output image. The noise data is self-consistent, meaning that the data obtained by mixing and re-splitting the noise data corresponding to the multiple initialized images is consistent with the noise data. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below.

[0018] Figure 1 This is a schematic flowchart of a slice image processing method according to an embodiment of this application;

[0019] Figure 2 This is a schematic diagram of the structure of a slice image processing apparatus according to an embodiment of this application;

[0020] Figure 3 This is a structural block diagram of a network device according to an embodiment of this application;

[0021] Figure 4 This is a structural block diagram of a user equipment according to one embodiment of this application. Detailed Implementation

[0022] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals identify the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0023] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms, while “a plurality” refers to two or more, and other quantifiers are similarly understood. It should be further understood that the word “comprising” as used in this application’s specification means the presence of the stated feature, integer, step, operation, element, and / or component, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connection or wireless coupling. The word “and / or” as used herein describes the relationship between related objects, indicating that three relationships can exist; for example, A and / or B can represent: A alone, A and B simultaneously, and B alone. The character “ / ” generally indicates that the preceding and following related objects are in an “or” relationship.

[0024] The proposed solution can be applied to nuclear magnetic resonance (NMR) image acquisition scenarios, reducing the number of NMR data acquisitions and improving sampling efficiency. Specifically, existing solutions, to avoid noise generated by simultaneously acquiring data from various slices, typically use NMR technology to scan each slice of the object separately, thereby acquiring data from each slice for reconstruction. This solution can acquire scan data containing information about overlapping and ambiguity of multiple slices of the object. After acquiring the scan data, it can be initialized into an initial image corresponding to the multiple slices of the object, and iterative processing is used to maintain image consistency during noise removal, thus obtaining a multi-slice image corresponding to the multiple slices of the object, facilitating subsequent 3D reconstruction and other processing.

[0025] Specifically, this application provides a method for processing sliced ​​images, such as... Figure 1 As shown, the method includes:

[0026] Step 102: Obtain scan data, which includes information corresponding to multiple slices of the scanned object.

[0027] Step 104: Initialize the scan data to obtain multiple initialized images corresponding to the multi-layer slices.

[0028] Step 106: Iteratively process multiple initial images to obtain multiple slice images. Each iteration includes analyzing the noise data of the input image and processing the input image based on the noise data to obtain the output image. The noise data is self-consistent. The self-consistency of the noise data means that the data obtained by mixing and re-splitting the noise data corresponding to multiple initial images is consistent with the noise data.

[0029] The proposed solution can be applied to nuclear magnetic resonance (NMR) image acquisition scenarios, reducing the number of NMR data acquisitions and improving sampling efficiency. Specifically, existing solutions, to avoid noise generated by simultaneously acquiring data from various slices, typically use NMR technology to scan each slice of the object separately, thereby acquiring data from each slice for reconstruction. This solution can acquire scan data containing information about overlapping and confusing layers of the object's slices. After acquiring the scan data, it can be initialized into an initial image corresponding to the multi-layer slices of the object, and iterative processing is used to maintain image consistency during noise removal, thus obtaining multi-layer slice images corresponding to the multi-layer slices of the object, facilitating subsequent 3D reconstruction and other processing. This solution utilizes simultaneous multi-slice scanning to acquire scan data from the multi-layer slices of the object for separation, denoising, and other processing, reducing the number of samplings, reducing scanning time, improving sampling efficiency, and enhancing reconstruction results. Rapid magnetic resonance imaging (MRI) technology, also known as multi-slice excitation technology or simultaneous multi-slice imaging, is a parallel imaging-based accelerated magnetic resonance imaging technique. In simultaneous multi-slice imaging, multiple slices are simultaneously excited by multi-frequency excitation pulses, thereby reducing scanning time and achieving rapid imaging. In fast magnetic resonance imaging methods, simultaneous multi-slice imaging is a type of inter-slice acceleration. Its biggest advantage over intra-slice acceleration methods is that the signal-to-noise ratio only increases by the square root of the number of slices.

[0030] This approach combines multi-layer excitation technology with a diffusion model to achieve better reconstruction results. The diffusion model, also known as a processing model, analyzes the noise contained in multiple input images for subsequent denoising. During the model training phase, this approach trains the processing model by iteratively adding noise to the training images. The added noise is self-consistent, so the diffusion model, after training, learns this self-consistent noise. During the initial image processing, the corresponding self-consistent noise is analyzed and denoised according to appropriate weights. After multiple iterations, multiple images corresponding to different slice locations are obtained.

[0031] Specifically, this solution can acquire scan data, which includes information corresponding to multiple slices of the scanned object; initialize the scan data to obtain multiple initialized images corresponding to the multiple slices; then, perform multiple iterative processing based on the multiple initialized images and a pre-trained processing model to obtain multiple slice images. Each iterative processing includes analyzing the noise data of the input image and processing the input image based on the noise data to obtain the output image. The noise data is self-consistent, meaning that the data obtained by mixing and re-splitting the noise data corresponding to the multiple initialized images is consistent with the noise data.

[0032] This scheme can pre-train a diffusion model (processing model). The processing model is used to analyze the noise contained in the input multiple images for subsequent denoising processing. Specifically, as an optional embodiment, the iterative processing based on multiple initial images to obtain multiple slice images includes: determining the time step of the current iteration and multiple input images, wherein the multiple input images are multiple output images obtained from the previous iteration or multiple initial images; inputting the multiple input images and the iteration time step into the pre-trained processing model to obtain self-consistent noise data, wherein the processing model is used to analyze the noise contained in the input multiple images; completing the current iteration based on the multiple input images and noise data to determine multiple output images; and performing the next round of iteration based on the multiple output images until multiple slice images are obtained.

[0033] This solution employs a diffusion model, a generative model used to generate data similar to the training data. The diffusion model works by continuously adding Gaussian noise to corrupt the training data, then learning to reverse this noise process to recover the data. After training, the diffusion model can be used to generate data by simply passing randomly sampled noise through a learned denoising process.

[0034] During the model training phase, this approach trains the diffusion model by iteratively adding noise to the training images. The added noise is self-consistent, so the diffusion model will learn the self-consistent noise after training. During the initial image processing, the corresponding self-consistent noise can be analyzed and denoised according to the corresponding weights. After multiple iterations, multiple images corresponding to different slice positions are obtained.

[0035] In this scheme, the process of generating sliced ​​images using a diffusion model involves splitting the input image and iteratively denoising it. The diffusion model training process involves continuously adding noise to the data to form training data, and then using the diffusion model to learn the noise in the image. Specifically, as an optional embodiment, the method further includes the step of training the processing model: acquiring multiple training images and iteratively training these multiple training images to determine multiple training noise images; wherein, each iteration of training includes analyzing the training noise data of the training input image, and processing the training input image based on the noise data to obtain the training output image; the training noise data conforms to self-consistency, which means that the data obtained by mixing and re-splitting the training noise data corresponding to multiple training input images is consistent with the training noise data. Specifically, as an optional embodiment, the iterative training of multiple training images includes: determining multiple training input images for the current iteration, wherein the multiple training input images are the training output data of the previous iteration or multiple training images; generating training noise data corresponding to the multiple training input images for the current iteration; fusing the multiple training input images with the training noise of the current iteration for the next iteration, until multiple training noise images are determined; and training the processing model based on the multiple training noise images, the multiple training images, and the training noise data of each iteration, until a trained training processing model is obtained.

[0036] This scheme utilizes a processing model to analyze self-consistent noise. However, using only this noise has limitations in the denoising process (e.g., small differences may prevent obtaining suitable slice images). Therefore, for better denoising, this scheme can also randomly generate noise and perform denoising processing according to corresponding weights. Specifically, as an optional embodiment, the step of completing the current iteration processing based on multiple input images and noise data to determine multiple output images includes: generating randomized noise and performing self-consistent processing on the randomized noise to obtain randomized noise data; and completing the current iteration processing based on multiple input images, noise data, and randomized noise data to determine multiple output images. This scheme can use noise data and randomized noise data for denoising processing, where the weights corresponding to the noise data and the weights corresponding to the randomized noise data can be correlated, such as the weight corresponding to the noise data being the square of the weight corresponding to the randomized noise data.

[0037] To avoid the difference between the output image and the original data exceeding a limit after multiple iterations, this scheme can also convert the input image into round data (k-space data) of the same type as the scan data. Then, the difference between the round data and the scan data is analyzed to process the input image for this iteration based on the amount of difference. Specifically, as an optional embodiment, the step of completing this iteration based on multiple input images, noise data, and random noise data to determine multiple output images includes: determining the round data corresponding to the input image and analyzing the difference data between the round data and the scan data; completing this iteration based on multiple input images, noise data, random noise data, and difference data to determine multiple output images. The difference data can be fused with the input image according to corresponding weights.

[0038] This scheme can also perform self-consistency processing on the input images and form fusion data according to the corresponding fusion ratio, so as to fuse the fusion data with the input images. Specifically, as an optional embodiment, multiple input images conform to self-consistency. Self-consistency of multiple input images means that the image obtained by re-splitting the multiple input images after mixing them conforms to the consistency between the multiple input images and the original multiple input images. The step of completing the current iteration processing based on multiple input images, noise data, random noise data, and difference data to determine multiple output images includes: determining the fusion ratio corresponding to the multiple input images to determine the fusion data corresponding to the multiple input images; and completing the current iteration processing based on multiple input images, noise data, random noise data, difference data, and fusion data to determine multiple output images. In a single iteration processing process, this scheme can perform comprehensive analysis based on noise data, random noise data, difference data, and fusion data, and then process the input image to form the input image for the next iteration. After multiple iterations, multiple slice images corresponding to multiple layers of slices are formed.

[0039] Based on the above embodiments, this application also provides a slice image processing apparatus, such as... Figure 2 As shown, the device includes:

[0040] The scan data acquisition module 202 is used to acquire scan data, which includes information corresponding to multiple slices of the scanned object.

[0041] The scan data initialization module 204 is used to initialize the scan data to obtain multiple initialized images corresponding to the multi-layer slices.

[0042] The slice image generation module 206 is used to perform iterative processing based on multiple initial images to obtain multiple slice images. Each iteration includes analyzing the noise data of the input image and processing the input image based on the noise data to obtain the output image. The noise data is self-consistent. The self-consistency of the noise data means that the data obtained by mixing and re-splitting the noise data corresponding to multiple initial images is consistent with the noise data.

[0043] The implementation methods of this application are similar to those of the above method embodiments. For specific implementation methods, please refer to the specific implementation methods of the above method embodiments, which will not be repeated here.

[0044] The proposed solution can be applied to nuclear magnetic resonance (NMR) image acquisition scenarios, reducing the number of NMR data acquisitions and improving sampling efficiency. Specifically, existing solutions, to avoid noise generated by simultaneously acquiring data from various slices, typically use NMR technology to scan each slice of the object separately, thereby acquiring data from each slice for reconstruction. This proposed solution can acquire scan data containing information about overlapping and ambiguity of multiple slices of the object. After acquiring the scan data, it can be initialized into an initial image corresponding to the multiple slices of the object, and iterative processing is used to maintain image consistency during noise removal, thus obtaining a multi-slice image corresponding to the multiple slices of the object, facilitating subsequent 3D reconstruction and other processing. This proposed solution utilizes simultaneous multi-slice scanning to acquire scan data of the multiple slices of the object for separation, denoising, and other processing, reducing the number of samplings and improving sampling efficiency. Specifically, this solution can acquire scan data, which includes information corresponding to multiple slices of the scanned object; initialize the scan data to obtain multiple initialized images corresponding to the multiple slices; then, perform multiple iterative processing based on the multiple initialized images to obtain multiple slice images. Each iterative processing includes analyzing the noise data of the input image based on a pre-trained processing model, and processing the input image based on the noise data to obtain the output image. The noise data is self-consistent, meaning that the data obtained by mixing and re-splitting the noise data corresponding to the multiple initialized images is consistent with the noise data.

[0045] It should be noted that the division of units and / or modules in the embodiments of this application is illustrative and only represents a logical functional division. In actual implementation, there may be other division methods. Furthermore, the functional units and / or modules in the various embodiments of this application can be integrated into one processing unit and / or module, or each unit and / or module can exist physically separately, or two or more units and / or modules can be integrated into one unit and / or module. The integrated units and / or modules described above can be implemented in hardware or as software functional units and / or modules.

[0046] If the integrated units and / or modules are implemented as software functional units and / or modules and sold or used as independent products, they can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0047] Furthermore, the data transmission apparatus and data transmission method provided in the above embodiments are based on the same application concept. Since the methods and apparatus solve problems in similar principles, the implementation of the apparatus and methods can refer to each other, and repeated parts will not be described again.

[0048] Figure 3 A structural block diagram of a network device is shown according to an exemplary embodiment.

[0049] like Figure 3 As shown, the network device 1100 includes at least: a processor 1110, a memory 1120, and a transceiver 1130.

[0050] The transceiver 1130 is used to receive and send data under the control of the processor 1110.

[0051] exist Figure 3In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits of one or more processors represented by processor 1110 and memory represented by memory 1120 together. The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 1130 can be multiple elements, including transmitters and receivers, providing units and / or modules for communicating with various other devices over transmission media, including wireless channels, wired channels, optical fibers, and other transmission media.

[0052] The processor 1110 is responsible for managing the bus architecture and general processing, and the memory 1120 can store the data used by the processor 1110 when performing operations.

[0053] Optionally, the processor 1110 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD). The processor 1110 may also adopt a multi-core architecture. The processor 1110 and the memory 1120 may also be physically separated.

[0054] The processor 1110 calls the computer program stored in the memory 1120 to execute any of the cell wireless network temporary identifier allocation methods provided in the above embodiments of this application according to the obtained executable instructions.

[0055] Figure 4 A structural block diagram of a user equipment is shown according to an exemplary embodiment.

[0056] like Figure 4 As shown, the user equipment 1300 includes at least: a processor 1310, a memory 1320, and a transceiver 1330.

[0057] The transceiver 1330 is used to receive and send data under the control of the processor 1310.

[0058] exist Figure 4In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits of one or more processors represented by processor 1310 and memory represented by memory 1320 together. The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 1330 can be multiple elements, including transmitters and receivers, providing units and / or modules for communicating with various other devices over a transmission medium, including wireless channels, wired channels, optical fibers, etc. For different user equipment, the user interface 1340 can also be an interface capable of connecting external or internal devices, including but not limited to keypads, displays, speakers, microphones, joysticks, etc.

[0059] The processor 1310 is responsible for managing the bus architecture and general processing, and the memory 1320 can store the data used by the processor 1310 when performing operations.

[0060] Optionally, the processor 1310 can be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a CPLD (Complex Programmable Logic Device). The processor 1310 can also adopt a multi-core architecture. The processor 1310 and the memory 1320 can also be physically separated.

[0061] The processor 1310 calls the computer program stored in the memory 1320 to execute any of the cell wireless network temporary identifier allocation methods provided in the above embodiments of this application according to the obtained executable instructions.

[0062] It should be noted that the apparatus provided in this application embodiment can implement all the method steps implemented in the above method embodiment and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.

[0063] Furthermore, this application provides a storage medium storing a computer program, which, when executed by a processor, implements the data transmission methods described in the above embodiments. The storage medium can be any available medium or data storage device accessible to the processor, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0064] This application provides a program product, such as an FPGA chip or a DSP chip, which includes executable instructions stored in a storage medium. A processor reads the executable instructions from the storage medium, causing the processor to execute the executable instructions to implement the data transmission methods described in the above embodiments.

[0065] The proposed solution can be applied to nuclear magnetic resonance (NMR) image acquisition scenarios, reducing the number of NMR data acquisitions and improving sampling efficiency. Specifically, existing solutions, to avoid noise generated by simultaneously acquiring data from various slices, typically use NMR technology to scan each slice of the object separately, thereby acquiring data from each slice for reconstruction. This solution can acquire scan data containing information about overlapping and ambiguity of multiple slices of the object. After acquiring the scan data, it can be initialized into an initial image corresponding to the multiple slices of the object, and iterative processing is used to maintain image consistency during noise removal, thus obtaining a multi-slice image corresponding to the multiple slices of the object, facilitating subsequent 3D reconstruction and other processing. This solution can acquire scan data of multiple slices of the object simultaneously using multi-slice scanning for separation, denoising, and other processing, reducing the number of samplings and improving sampling efficiency. Specifically, this solution can pre-train a processing model, which analyzes the noise contained in the input multiple images for subsequent denoising processing. This scheme can acquire scan data, which includes information corresponding to multiple slices of the scanned object; initialize the scan data to obtain multiple initialized images corresponding to the multiple slices; then, perform multiple iterative processing based on the multiple initialized images and a pre-trained processing model to obtain multiple slice images. Each iteration includes analyzing the noise data of the input image based on the pre-trained processing model, and processing the input image based on the noise data to obtain the output image. The noise data is self-consistent, meaning that the data obtained by mixing and re-splitting the noise data corresponding to the multiple initialized images is consistent with the noise data.

[0066] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0067] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-executable instructions. These computer-executable instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0068] These processor-executable instructions may also be stored in a processor-readable memory that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the processor-readable memory produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0069] These processors can execute instructions that can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0070] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0071] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for processing sliced ​​images, characterized in that, The method includes: Acquire scan data, which includes information about the multi-layer slices of the scanned object; The scanned data is initialized to obtain multiple initialized images corresponding to the multi-layer slices; Multiple initial images are iteratively processed to obtain multiple slice images. Each iteration includes analyzing the noise data of the input image and processing the input image based on the noise data to obtain the output image. The noise data is self-consistent, meaning that the data obtained by mixing and re-splitting the noise data corresponding to multiple initial images is consistent with the noise data. The step of iteratively processing multiple initial images to obtain multiple slice images includes: determining the time step and multiple input images for the current iteration, wherein the multiple input images are multiple output images or multiple initial images obtained from the previous iteration; inputting the multiple input images and the iteration time step into a pre-trained processing model to obtain self-consistent noise data, wherein the processing model is used to analyze the noise contained in the images based on the input multiple images; completing the current iteration based on the multiple input images and noise data to determine multiple output images; and determining the next round of iteration based on the multiple input images until multiple slice images are obtained.

2. The method as described in claim 1, characterized in that, The iterative processing, based on multiple input images and noise data, determines multiple output images, including: Randomized noise is generated, and self-consistency processing is performed on the randomized noise to obtain random noise data; The current iteration is completed based on multiple input images, noise data, and random noise data, and multiple output images are determined.

3. The method as described in claim 2, characterized in that, The iterative processing is completed based on multiple input images, noise data, and random noise data to determine multiple output images, including: Determine the round data corresponding to the input image and analyze the differences between the round data and the scan data; Based on multiple input images, noise data, random noise data, and difference data, this iterative processing is completed to determine multiple output images.

4. The method as described in claim 3, characterized in that, Multiple input images conform to self-consistency. Self-consistency of multiple input images means that the image obtained by mixing and splitting multiple input images conforms to the consistency of the multiple input images. The iterative processing is completed based on multiple input images, noise data, random noise data, and difference data to determine multiple output images, including: Based on multiple input images, determine the fusion ratio corresponding to the multiple input images in order to determine the fusion data corresponding to the multiple input images; Based on multiple input images, noise data, random noise data, difference data, and fused data, this iterative processing is completed to determine multiple output images.

5. The method as described in claim 1, characterized in that, The method also includes the step of training a processing model: Acquire multiple training images and iteratively train on these images to identify multiple training noise images; Each iteration of training includes analyzing the training noise data of the training input image and processing the training input image based on the noise data to obtain the training output image. The training noise data is self-consistent, meaning that the data obtained by mixing and re-splitting the training noise data corresponding to multiple training input images is consistent with the training noise data.

6. The method as described in claim 5, characterized in that, The iterative training of multiple training images includes: The multiple training input images for this iteration are determined. These multiple training input images are either the training output data from the previous iteration or multiple training images. Generate training noise data corresponding to multiple training input images for this iteration round; Multiple training input images are fused with the training noise from the current iteration for the next iteration of training, until multiple training noise images are determined. The processing model is trained based on multiple training noisy images, multiple training images, and training noisy data from each iteration until a well-trained processing model is obtained.

7. A slice image processing apparatus, characterized in that, The device includes: The scan data acquisition module is used to acquire scan data, which includes information corresponding to multiple slices of the scanned object. The scan data initialization module is used to initialize the scan data and obtain multiple initialized images corresponding to the multi-layer slices; The slice image generation module is used to perform iterative processing based on multiple initial images to obtain multiple slice images. Each iteration includes analyzing the noise data of the input image and processing the input image based on the noise data to obtain the output image. The noise data is self-consistent. The self-consistency of the noise data means that the data obtained by mixing and re-splitting the noise data corresponding to multiple initial images is consistent with the noise data. The step of iteratively processing multiple initial images to obtain multiple slice images includes: determining the time step and multiple input images for the current iteration, wherein the multiple input images are multiple output images or multiple initial images obtained from the previous iteration; inputting the multiple input images and the iteration time step into a pre-trained processing model to obtain self-consistent noise data, wherein the processing model is used to analyze the noise contained in the images based on the input multiple images; completing the current iteration based on the multiple input images and noise data to determine multiple output images; and determining the next round of iteration based on the multiple input images until multiple slice images are obtained.

8. A network device, characterized in that, include: The system includes a memory, a transceiver, and a processor; wherein the memory is used to store computer programs; and the transceiver is used to send and receive data under the control of the processor. The processor is configured to read a computer program from the memory and execute the method as described in any one of claims 1-6.

9. A storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1-6.

Images