Virtual enhancement processing method and device based on MTT sequence
The virtual enhanced images are generated through multi-task magnetic resonance imaging technology and image processing model, which solves the problem of patients' intolerance to contrast agents and realizes accurate judgment of myocardial injury images.
Patent Information
- Application Number
- CN202411151392.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-21
- Publication Date
- 2025-01-03
AI Technical Summary
Many patients are intolerable to the contrast agent required for cardiac magnetic resonance delay enhancement (LGE) images, resulting in the inability to obtain accurate images of myocardial injury, affecting the diagnosis.
Multi-sequence images are acquired based on multi-task magnetic resonance imaging technology (MTT) and virtual enhancement processing is performed through image processing model to generate virtual enhancement images consistent with LGE images, which are used to indicate myocardial vitality and infarction status and avoid the use of contrast agents. .
It realizes the acquisition of virtual enhancement images that are the same or similar to LGE images without contrast agent, which solves the problem that patients cannot obtain myocardial injury images and improves the accuracy of myocardial injury judgment.
Smart Images
Figure CN119251060A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the technical field of image processing, and in particular to a virtual enhancement processing method and device based on an MTT sequence. Background Art
[0002] Myocardial injury refers to the damage to myocardial cells due to various reasons, and even the occurrence of necrosis. Mild myocardial injury may not have any symptoms, or only mild symptoms such as chest tightness and shortness of breath. Severe myocardial injury can lead to chest pain, heart failure, arrhythmia, and even sudden death.
[0003] In clinical practice, cardiac magnetic resonance delayed enhancement (LGE) images are often needed to determine myocardial damage. If LGE images are to be scanned for patients, contrast agents need to be used on the patients. However, many patients are intolerant to contrast agents, so these patients cannot use contrast agents. Therefore, how to obtain LGE images of these patients is an urgent problem to be solved in this field. Summary of the invention
[0004] In view of this, the purpose of the present disclosure is to propose a virtual enhancement processing method and device based on MTT sequence, and the present disclosure can solve the existing problems in a targeted manner.
[0005] Based on the above-mentioned purpose, in the first aspect, the present disclosure proposes a virtual enhancement processing method based on the MTT sequence, comprising: acquiring a multi-sequence image, wherein the multi-sequence image is used to indicate the damage of the myocardial tissue and is reconstructed based on the multi-task magnetic resonance imaging technology; calling the image processing model, performing post-processing analysis on the multi-sequence image, and generating a virtual enhanced VNE image, wherein the consistency between the VNE image and the LGE image corresponding to the myocardial tissue is greater than or equal to a preset consistency threshold, and the VNE image is used to indicate at least one of myocardial vitality and infarction status.
[0006] In the second aspect, a virtual enhancement processing device based on the MTT sequence is also provided, characterized in that it includes: an acquisition unit, configured to acquire multi-sequence images, wherein the multi-sequence images are used to indicate damage to myocardial tissue and are reconstructed based on multi-task magnetic resonance imaging technology; a processing unit, configured to call an image processing model, perform post-processing analysis on the multi-sequence images, and generate a virtual enhanced VNE image, wherein the consistency between the virtual enhanced image and the LGE image corresponding to the myocardial tissue is greater than or equal to a preset consistency threshold, and the VNE image is used to indicate at least one of myocardial vitality and infarction status.
[0007] According to a third aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect.
[0008] According to a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and the program is executed by a processor to implement any method described in the first aspect.
[0009] In general, the present disclosure has at least the following beneficial effects: by performing virtual enhancement processing on the sequence images, it is possible to obtain virtual enhanced images that are the same as or similar to the LGE images without using contrast agents on the patient's myocardial tissue, thereby solving the problem that many patients are intolerant to contrast agents and cannot obtain LGE images, which affects the judgment of myocardial damage. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the multiple drawings represent the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings only depict some embodiments disclosed in accordance with the present disclosure and should not be regarded as limiting the scope of the present disclosure.
[0011] Figure 1 A flowchart of a virtual enhancement processing method based on an MTT sequence according to an embodiment of the present disclosure is shown;
[0012] Figure 2 Another flow chart of a virtual enhancement processing method based on an MTT sequence according to an embodiment of the present disclosure is shown;
[0013] Figure 3 A flowchart of virtual enhancement processing and model training based on MTT sequences according to an embodiment of the present disclosure is shown;
[0014] Figure 4 A schematic diagram of a virtual enhancement processing device based on an MTT sequence according to an embodiment of the present disclosure is shown;
[0015] Figure 5 A schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure is shown;
[0016] Figure 6 A schematic diagram of a storage medium provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION
[0017] The present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.
[0018] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0019] Figure 1 The virtual enhancement processing method based on MTT sequence of the present disclosure is shown. In an embodiment of the present disclosure, the method includes:
[0020] Step S101 : acquiring a multi-sequence image, wherein the multi-sequence image is used to indicate damage to myocardial tissue and is reconstructed based on a multi-task magnetic resonance imaging technology.
[0021] In this embodiment, when there are multiple sequence images, i.e., multiple sequence images, each of the multiple sequence images indicates a different type of myocardial injury. The multi-task tissue feature quantitative imaging technology MTT is used to determine the sequence image of the myocardial tissue. Specifically, the multi-task magnetic resonance imaging technology (MR Multitasking, MTT) can be used to scan the myocardial tissue to obtain the above sequence images.
[0022] The execution subject of the MTT sequence-based virtual enhancement processing method can adopt multi-task tissue feature quantitative imaging technology to obtain sequence images through a cardiac magnetic resonance scanning device, or can also obtain sequence images from other electronic devices.
[0023] Specifically, the sequence of images may include a cine sequence (Cardiac cine imaging) and a T1 sequence. A cine sequence is an image sequence obtained by performing a nuclear magnetic resonance scan on myocardial tissue. A T1 sequence is an image sequence obtained by performing a pulse inversion on the result of a nuclear magnetic resonance scan.
[0024] In addition, the sequence images may also include other images obtained by further processing the images obtained by the nuclear magnetic resonance working scan, such as T1 rho, T2 or T2*mapping images and Cine images.
[0025] Step S102, calling the image processing model, performing post-processing analysis on the multi-series images, and generating a virtual enhanced VNE image, wherein the consistency between the VNE image and the cardiac magnetic resonance delayed enhancement LGE image corresponding to the myocardial tissue is greater than or equal to a preset consistency threshold, and the VNE image is used to indicate at least one of myocardial vitality and infarction status.
[0026] In this embodiment, the execution subject may call the image processing model to process the sequence images, so as to perform post-processing analysis on the multiple sequence images, namely, virtual native enhancement (VNE) processing, thereby obtaining a virtual enhanced image.
[0027] The image processing model is a deep neural network, such as a generative adversarial network or a convolutional neural network. The goal of the virtual enhancement processing is to make the consistency between the obtained virtual enhanced image and the LGE image greater than or equal to a preset consistency threshold.
[0028] This embodiment can obtain virtual enhanced images that are the same as or similar to LGE images by virtually enhancing the sequence images without using contrast agents on the patient's myocardial tissue, thereby solving the problem that many patients are intolerant to contrast agents and cannot obtain LGE images, which affects the judgment of myocardial damage.
[0029] In some optional implementations of any embodiment of the present application, the multi-sequence image includes a first image, a second image and a third image, the first image is used to indicate image contrast, the second image is used to indicate signal changes in myocardial tissue characteristics, and the third image is used to indicate additional wall motion information and myocardial boundaries.
[0030] Among these implementations, T1 rho, T1, T2, and T2*mapping images provide image contrast and signal changes of myocardial tissue characteristics, and Cine images provide additional wall motion information and clearer myocardial boundaries.
[0031] The magnetic resonance scanner is subjected to flip pulses, T2 preparation pulses and T1ρ preparation pulses to obtain T1 contrast, T2 contrast and T1ρ contrast respectively, and a multi-echo image is reconstructed by the T1 contrast, T2 contrast and T1ρ contrast to obtain the multi-sequence image.
[0032] In some optional implementations of any embodiment of the present application, the method further includes: performing semi-automatic myocardial lesion quantification on the VNE image and the LGE image to qualitatively and / or quantitatively determine the accuracy of the VNE image and the LGE image.
[0033] Optionally, the semi-automatic myocardial lesion quantification of the VNE image and the LGE image to qualitatively and / or quantitatively determine the accuracy of the VNE image and the LGE image includes: automatically initializing the epicardial and endocardial left ventricular contours of the VNE image and the LGE image to obtain an initialized image; acquiring an adjusted image of the initialized image; adding a distal reference region of interest to the VNE image, and using a half-maximum full width method to adaptively calculate the lesion load of the myocardial tissue.
[0034] Semiautomated myocardial lesion quantification by VNE and LGE was performed as follows: epicardial and endocardial left ventricular contours were automatically initialized and manually adjusted for all images using regions of interest (ROIs) by an experienced operator. A distal reference ROI without LGE was added, and lesion burden was calculated using an adaptation of the full width at half maximum (FWHM) method.
[0035] Through the evaluation, we can further determine whether the generated VNE image is accurate and whether it can be comparable to the accuracy of the accurate LGE image.
[0036] Figure 2 The training method of the image processing model according to the embodiment of the present disclosure is shown. The image processing model is a generative adversarial network, and the generative adversarial network includes at least one generator and at least one discriminator. Figure 2 As shown, the steps of the training method include:
[0037] Step S201: input the sequence image samples into the generator to generate an initial virtual enhanced image.
[0038] In this embodiment, the generative adversarial network may be a conditional generative adversarial network (cGAN).
[0039] Step S202: using the discriminator, determining whether the initial virtual enhanced image is an LGE image.
[0040] In this embodiment, the execution subject may use the at least one discriminator in various ways to determine whether the initial virtual enhanced image is a virtual enhanced image obtained through virtual enhancement processing. For example, it may be determined whether the image consistency between the initial virtual enhanced image and the LGE image sample is greater than a preset threshold. The LGE image sample and the sequence image sample correspond to the same myocardial tissue.
[0041] Step S203: training the generator according to the initial virtual enhanced image and the LGE image sample, and training the generator and the discriminator according to the judgment result and the reference judgment result.
[0042] In this embodiment, the above-mentioned execution entity can determine the consistency between the initial virtual enhanced image and the LGE image sample, and determine the loss value of the generator based on the consistency of the sequence image samples marked in the training samples, and train the generator through the loss value.
[0043] In the training sample set, there are reference judgment results corresponding to the sequence images, that is, it can indicate that the virtual enhanced image output by the generator is a virtual enhanced image obtained through virtual enhancement processing, rather than other images such as LGE images.
[0044] In some optional implementations of this embodiment, the at least one discriminator includes a first discriminator; and using the discriminator to determine whether the initial virtual enhanced image is an LGE image includes: inputting the initial virtual enhanced image into the first discriminator to determine whether the initial virtual enhanced image is a VNE image generated by the post-processing analysis, or an LGE image, to obtain a first judgment result, and the judgment result includes the first judgment result.
[0045] The execution subject may determine whether the initial virtual enhanced image is a virtual enhanced image obtained through post-processing analysis or a LGE image through the first discriminator. The determination result may include only the first determination result, and may also include other determination results.
[0046] In some optional implementations of this embodiment, the at least one discriminator also includes a second discriminator; the initial virtual enhanced image is input into the second discriminator to judge whether the initial virtual enhanced image is a VNE image or a Cine image generated by the post-processing analysis through the clarity of the initial virtual enhanced image, and obtain a second judgment result, which includes the second judgment result.
[0047] The execution subject may determine whether the initial virtual enhanced image is a virtual enhanced image obtained through virtual enhancement processing or a Cine image through the second discriminator. The judgment result may include only the first judgment result, and may also include other judgment results, so that the trained generator matches the high-quality Cine image in terms of image clarity.
[0048] The judgment result may include the second judgment result and the second judgment result, and may also include the second judgment result and other judgment results, such as the first judgment result and the second judgment result.
[0049] In some optional application scenarios of these implementation methods, a first loss value corresponding to a pixel consistency loss function is determined through pixel information of both the initial virtual enhanced image and the LGE image sample, and a second loss value corresponding to a perceptual consistency loss function is determined through feature information of the two; and the image processing model is trained based on the first loss value and the second loss value.
[0050] In these application scenarios, the above-mentioned execution subject can determine the pixel information of the initial virtual enhanced image and the pixel information of the LGE image sample. And according to these two pixel information, the pixel consistency loss function is input to obtain the first loss value. The pixel information is the parameter value that needs to be filled in the pixel consistency function, such as the pixel value. Through training with this loss function, it can be ensured that the virtual enhanced image obtained after training is consistent with the real LGE image in each pixel.
[0051] The execution subject can determine the feature information of the initial virtual enhanced image and the feature information of the LGE image sample. According to the two feature information, the perceptual consistency loss function is input to obtain a second loss value. By training with the perceptual consistency loss function, it can be ensured that the features of the virtual enhanced image obtained after training are consistent with those of the real LGE image.
[0052] The execution subject may train the image processing model according to the first loss value and the second loss value in various ways. For example, the sum of the first loss value and the second loss value may be determined as the total loss value, and the image processing model may be trained using the total loss value. Alternatively, the first loss value and the second loss value may be input into a preset formula, and the total loss value corresponding to the formula may be obtained.
[0053] In some optional implementations of any embodiment of the present application, the generator includes multiple U-net structures in parallel, and the U-net structure includes an encoder and a decoder; the calling of the image processing model to perform virtual enhancement processing on the sequence image to obtain a virtual enhanced image includes: extracting features of the sequence image through the convolution layer and the pooling layer in the encoder to obtain multi-scale features; combining the multi-scale features through the decoder, and generating a feature vector for the combination result through the deconvolution layer in the decoder, cascading each feature vector, and fusing the cascaded result to obtain the virtual enhanced image.
[0054] In these implementations, the generator consists of 2-5 parallel U-net networks. The U-net network adopts an encoder-decoder architecture. The encoder reduces the image size and extracts features through convolutional layers and pooling layers. The decoder combines multi-scale features and upsamples the image through a deconvolution layer to generate a feature vector. The feature vectors generated by multiple parallel U-Nets are cascaded and input into the fusion module of the image processing model for multimodal information fusion to generate a virtual enhanced image. Different sequence images, such as T1 and movie sequences, are images of different modalities.
[0055] The present disclosure also provides a flowchart of virtual enhancement processing and model training based on MTT sequences.
[0056] The embodiment of the present disclosure provides a virtual enhancement processing device based on an MTT sequence, and the device is used to execute the virtual enhancement processing method based on an MTT sequence described in the above embodiment, such as Figure 4 As shown, the device includes: an acquisition unit 401, configured to acquire a multi-sequence image, wherein the multi-sequence image is used to indicate damage to myocardial tissue and is reconstructed based on a multi-task magnetic resonance imaging technology; a processing unit 402, configured to call an image processing model, perform post-processing analysis on the multi-sequence image, and generate a virtual enhanced VNE image, wherein the consistency between the virtual enhanced image and the LGE image corresponding to the myocardial tissue is greater than or equal to a preset consistency threshold.
[0057] The MTT sequence-based virtual enhancement processing device provided in the above-mentioned embodiment of the present disclosure and the MTT sequence-based virtual enhancement processing method provided in the embodiment of the present disclosure are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.
[0058] The embodiment of the present disclosure also provides an electronic device corresponding to the virtual enhancement processing method based on the MTT sequence provided in the above embodiment, so as to execute the above virtual enhancement processing method based on the MTT sequence. The embodiment of the present disclosure is not limited.
[0059] Please refer to Figure 5 , which shows a schematic diagram of an electronic device provided by some embodiments of the present disclosure. Figure 5 As shown, the electronic device 50 includes: a processor 500, a memory 501, a bus 502 and a communication interface 503, and the processor 500, the communication interface 503 and the memory 501 are connected via the bus 502; the memory 501 stores a computer program that can be run on the processor 500, and the processor 500 executes the method provided in any of the aforementioned embodiments of the present disclosure when running the computer program.
[0060] The memory 501 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 503 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.
[0061] The bus 502 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 501 is used to store a program, and the processor 500 executes the program after receiving an execution instruction. The virtual enhancement processing method based on the MTT sequence disclosed in any implementation of the aforementioned embodiment of the present disclosure may be applied to the processor 500, or implemented by the processor 500.
[0062] The processor 500 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit in the processor 500 or the instruction in the form of software. The above processor 500 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present disclosure can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in conjunction with the embodiments of the present disclosure can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 501, and the processor 500 reads the information in the memory 501 and completes the steps of the above method in combination with its hardware.
[0063] The electronic device provided by the embodiment of the present disclosure and the virtual enhancement processing method based on the MTT sequence provided by the embodiment of the present disclosure are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by them.
[0064] The present disclosure also provides a computer-readable storage medium corresponding to the virtual enhancement processing method based on the MTT sequence provided in the above embodiment. Figure 6 The computer-readable storage medium shown is a CD 60 on which a computer program (ie, a program product) is stored. When the computer program is run by a processor, the virtual enhancement processing method based on the MTT sequence provided in any of the aforementioned embodiments will be executed.
[0065] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.
[0066] The computer-readable storage medium provided by the above-mentioned embodiment of the present disclosure and the virtual enhancement processing method based on the MTT sequence provided by the embodiment of the present disclosure are based on the same inventive concept, and have the same beneficial effects as the method adopted, run or implemented by the application program stored therein.
[0067] It should be noted that:
[0068] In the above text, the term "comprises", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be noted that the scope of the method and device in the embodiments of the present disclosure is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.
[0069] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present disclosure, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present disclosure.
[0070] The embodiments of the present disclosure are described above in conjunction with the accompanying drawings, which are only specific implementation modes of the present disclosure. However, the present disclosure is not limited to the above-mentioned specific implementation modes. The above-mentioned specific implementation modes are merely illustrative and not restrictive. Under the guidance of the present disclosure, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present disclosure and the claims, all of which are within the protection of the present disclosure.
Claims
1. A virtual enhancement processing method based on MTT sequence, characterized in that: include: Acquiring a multi-sequence image, wherein the multi-sequence image is used to indicate damage to myocardial tissue and is reconstructed based on a multi-task magnetic resonance imaging technology; An image processing model is called to perform post-processing analysis on the multiple sequence images to generate a virtual enhanced VNE image, wherein the consistency between the VNE image and the cardiac magnetic resonance delayed enhancement LGE image corresponding to the myocardial tissue is greater than or equal to a preset consistency threshold, and the VNE image is used to indicate at least one of myocardial vitality and infarction status.
2. The method according to claim 1, characterized in that The multi-sequence image includes a first image, a second image and a third image, wherein the first image is used to indicate image contrast, the second image is used to indicate signal changes in myocardial tissue characteristics, and the third image is used to indicate additional wall motion information and myocardial boundaries.
3. The method according to claim 1, characterized in that The method further comprises: Semi-automatic myocardial lesion quantification is performed on the VNE image and the LGE image to qualitatively and / or quantitatively determine the accuracy of the VNE image and the LGE image.
4. The method according to claim 3, characterized in that The semi-automatic myocardial lesion quantification of the VNE image and the LGE image to qualitatively and / or quantitatively determine the accuracy of the VNE image and the LGE image comprises: For the VNE image and the LGE image, automatically initializing the epicardial and endocardial left ventricular contours to obtain an initialized image; Acquire an adjusted image of the initialization image; A distal reference region of interest was added to the VNE image, and the lesion burden of the myocardial tissue was calculated using the full width at half maximum method.
5. The method according to claim 1, characterized in that The image processing model is a generative adversarial network, which includes at least one generator and at least one discriminator; The training step of the image processing model includes: Inputting the sequence image samples into the generator to generate an initial virtual enhanced image; Using the discriminator, determining whether the initial virtual enhanced image is an LGE image; The generator is trained according to the initial virtual enhanced image and the LGE image sample, and the generator and the discriminator are trained according to the judgment result and the reference judgment result.
6. The method according to claim 5, characterized in that The at least one discriminator includes a first discriminator; and the using the discriminator to determine whether the initial virtual enhanced image is an LGE image includes: The initial virtual enhanced image is input into the first discriminator to determine whether the initial virtual enhanced image is a VNE image or a LGE image generated through the post-processing analysis, and a first judgment result is obtained, which includes the first judgment result.
7. The method according to claim 6, characterized in that The at least one discriminator further includes a second discriminator; and the using the discriminator to determine whether the initial virtual enhanced image is an LGE image includes: The initial virtual enhanced image is input into the second discriminator to judge whether the initial virtual enhanced image is a VNE image or a Cine image generated by the post-processing analysis through the clarity of the initial virtual enhanced image, so as to obtain a second judgment result, which includes the second judgment result.
8. The method according to claim 7, characterized in that The step of training the generator according to the initial virtual enhanced image and the LGE image sample comprises: Determine a first loss value corresponding to a pixel consistency loss function through pixel information of the initial virtual enhanced image and the LGE image sample, and determine a second loss value corresponding to a perceptual consistency loss function through feature information of the two; The image processing model is trained according to the first loss value and the second loss value.
9. The method according to claim 5, characterized in that The generator includes a plurality of U-net structures connected in parallel, and the U-net structure includes an encoder and a decoder; The calling of the image processing model to perform post-processing analysis on the multiple sequence images to generate a virtual enhanced VNE image includes: Extracting features from the sequence images through the convolutional layer and the pooling layer in the encoder to obtain multi-scale features; The multi-scale features are combined by the decoder, and a feature vector is generated from the combination result by a deconvolution layer in the decoder, each feature vector is cascaded, and the cascaded result is fused to obtain the virtual enhanced image.
10. A virtual enhancement processing device based on MTT sequence, characterized in that: include: An acquisition unit is configured to acquire a multi-series image, wherein the multi-series image is used to indicate damage to myocardial tissue and is reconstructed based on a multi-task magnetic resonance imaging technology; The processing unit is configured to call an image processing model, perform post-processing analysis on the multiple sequence images, and generate a virtual enhanced VNE image, wherein the consistency between the VNE image and the cardiac magnetic resonance delayed enhancement LGE image corresponding to the myocardial tissue is greater than or equal to a preset consistency threshold, and the VNE image is used to indicate at least one of myocardial vitality and infarction status.