MRI (Magnetic Resonance Imaging) mismatching scoring method, device, equipment, medium and product

Abstract

The invention discloses an MRI mismatch scoring method, device, equipment, medium and product, and relates to the field of wisdom medical treatment, and the method comprises the steps: registering a low-field magnetic resonance cerebral apoplexy identification sequence image to a DWI image, and enabling a focus segmentation result on the DWI image to act on the registered low-field magnetic resonance cerebral apoplexy identification sequence image, and obtaining an area corresponding to a DWI high signal on the registered low-field magnetic resonance cerebral apoplexy identification sequence image, calculating an image intensity mean value of the area, and comparing the image intensity mean value of the area with the intensity mean value in the whole brain of the registered low-field magnetic resonance cerebral apoplexy identification sequence image to obtain a mismatching ratio. According to the method, DWI-low-field magnetic resonance cerebral apoplexy identification sequence mismatching identification can be carried out based on the DWI and the low-field magnetic resonance cerebral apoplexy identification sequence, so that the MRI mismatching scoring efficiency and the MRI mismatching scoring result accuracy are improved.

Description

Technical Field

[0001] The present application relates to the field of smart medical technology, and in particular to an MRI mismatch scoring method, device, equipment, medium and product. Background Art

[0002] Cerebrovascular diseases, including acute cerebral hemorrhage and cerebral infarction, are one of the leading causes of death and disability worldwide. Accurate and rapid identification of acute cerebral hemorrhage and cerebral infarction is crucial for timely treatment and improved prognosis. Acute cerebral hemorrhage refers to a disease in which cerebral blood vessels rupture, causing blood to enter the brain tissue and causing brain tissue damage. Cerebral infarction is a condition in which cerebral blood vessels are blocked, causing ischemia and hypoxia of the brain tissue, causing brain tissue damage. Although both involve problems with brain blood circulation, their etiology and pathophysiological mechanisms are different, and the treatment decisions are completely different: cerebral hemorrhage requires blood pressure control and reduction of cerebral edema; for the acute stage of cerebral infarction, thrombolysis and mechanical thrombectomy can be used, and antiplatelet aggregation can be performed later to improve circulation treatment. Therefore, being able to quickly identify cerebral hemorrhage and cerebral infarction in the hyperacute stage will be very beneficial for patients' treatment plan decisions.

[0003] The differential diagnosis between cerebral hemorrhage and cerebral infarction includes symptoms and imaging examination results. The usual differentiation method is as follows: cerebral hemorrhage usually occurs during activities, manifested as sudden headache, nausea, vomiting, and then combined with the clinical characteristics of hemiplegia, aphasia, and coma, it can basically be determined to be cerebral hemorrhage. Cerebral infarction usually occurs in a quiet state, and also manifests as hemiplegia, aphasia, and even coma. The most critical differentiation is that once this happens, if high density is seen on computed tomography (CT), it is determined to be cerebral hemorrhage. On the other hand, imaging examinations also play an important role in differential diagnosis. Commonly used imaging examinations include cranial CT and magnetic resonance imaging (MRI), which can help doctors determine the cause and type of disease by observing the condition of brain blood vessels and the degree of brain tissue damage. In addition, special examination methods such as angiography can be used to further clarify the diagnosis. Clinical scoring tools are an auxiliary diagnostic method that can help doctors more accurately judge the patient's condition. Commonly used clinical scoring tools include the Glasgow Coma Scale (GCS), the National Institute of Health Stroke Scale (NIHSS), etc., which can provide important reference for doctors' diagnosis by evaluating the patient's consciousness state, limb movement and other indicators.

[0004] For traditional imaging identification methods, CT has high sensitivity in detecting acute cerebral hemorrhage, but low sensitivity in detecting acute cerebral infarction. This limitation may lead to misdiagnosis or missed diagnosis, delay treatment, and thus affect the patient's prognosis. MRI has high sensitivity in detecting acute cerebral infarction, especially DWI is the most sensitive. Although MRI sequences can provide detailed brain structure information, identification cannot be achieved through certain imaging features in a certain MRI sequence. It is usually necessary to compare, observe and analyze multiple MRI sequences, and finally draw conclusions, which is very dependent on the doctor's experience. Therefore, the sensitivity and specificity of MRI sequences in hyperacute lesions are limited. This deficiency causes doctors to rely on experience and subjective judgment, increasing the risk of misdiagnosis.

[0005] MRI mismatch score is to evaluate the infarct core and ischemic penumbra of ischemic stroke patients in MRI through multiple imaging techniques, and then calculate the mismatch volume and mismatch ratio (i.e., mismatch ratio) to guide clinical treatment. MRI mismatch score is mainly based on imaging technology, such as DWI-PWI mismatch identification (i.e., DWI-PWI mismatch score) based on diffusion weighted imaging (DWI) and perfusion weighted imaging (PWI), or DWI-FLAIR mismatch identification based on DWI and fluid attenuated inversion recovery (FLAIR), or DWI-ASL mismatch identification based on DWI and arterial spin labeling (ASL), etc. In these mismatch identifications, such as DWI-PWI mismatch identification, the current mainstream methods are mainly based on traditional imaging analysis and mainly rely on doctors' interpretation of DWI-PWI. There are problems of low efficiency of MRI mismatch scoring and low accuracy of MRI mismatch scoring results. There are many MRI scanning sequences, including not only PWI, DWI, FLAIR, and ASL sequences, but also low-field magnetic resonance stroke identification sequences (i.e., low-field magnetic resonance stroke identification sequences). Like these sequences, such as DWI, low-field magnetic resonance stroke identification sequences are also scanned by magnetic resonance. DWI and low-field magnetic resonance stroke identification sequences can be quickly scanned within a few minutes using low-field mobile magnetic resonance equipment (such as the mobile head and neck magnetic resonance system ACUTAElfin manufactured by Foshan Regato Medical Technology Co., Ltd.). Compared with sequences such as PWI, FLAIR, and ASL, low-field magnetic resonance stroke identification sequences can quickly and accurately judge hemorrhagic stroke. Based on this, there is an urgent need in the art for an MRI mismatch scoring method for DWI-low-field magnetic resonance stroke identification sequence mismatch identification based on DWI and low-field magnetic resonance stroke identification sequences to improve the efficiency of MRI mismatch scoring and the accuracy of MRI mismatch scoring results.

[0006] However, the existing mismatch identification algorithms are mainly used for the evaluation between DWI and PWI, DWI and FLAIR, and DWI and ASL. There is no algorithm for the mismatch of DWI and low-field magnetic resonance stroke identification sequence signals, and the applicability and effect of the existing algorithms are limited. This blank limitation leads to the fact that in the identification of hyperacute cerebral hemorrhage and cerebral infarction, if other current image mismatch models (including DWI-PWI mismatch identification models and mismatch models of other different modal images, such as DWI-FLAIR, etc.) are directly applied, the effect is not ideal, because the current image algorithm can still basically only process images of known modalities during training. For low-field magnetic resonance stroke identification sequence images, which are unseen modalities, the algorithm generally needs to be redesigned and experimented. In view of the above-mentioned gaps in the prior art, the art urgently needs an algorithm that can perform DWI-low-field magnetic resonance stroke identification sequence mismatch identification based on DWI and low-field magnetic resonance stroke identification sequences. Summary of the invention

[0007] The purpose of this application is to provide an MRI mismatch scoring method, device, equipment, medium and product, which can perform DWI-low-field MRI stroke identification sequence mismatch identification based on DWI and low-field MRI stroke identification sequences to improve the efficiency of MRI mismatch scoring and the accuracy of MRI mismatch scoring results.

[0008] To achieve the above objectives, this application provides the following solutions.

[0009] In a first aspect, the present application provides an MRI mismatch scoring method, which comprises: obtaining a DWI image and a low-field magnetic resonance stroke identification sequence image of a target object; segmenting the DWI image using a DWI lesion segmentation model to obtain a lesion segmentation result on the DWI image; the DWI lesion segmentation model is obtained by training a neural network based on a U-Net framework using a DWI acute ischemic lesion data set; the lesion segmentation result is a binary mask; performing skull stripping on the DWI image and the low-field magnetic resonance stroke identification sequence image, respectively, to obtain a DWI image after skull stripping and a low-field magnetic resonance stroke identification sequence image after skull stripping; registering the low-field magnetic resonance stroke identification sequence image after skull stripping to the low-field magnetic resonance stroke identification sequence image after skull stripping. on the DWI images, to obtain the registered low-field magnetic resonance stroke identification sequence images; the lesion segmentation result is applied to the registered low-field magnetic resonance stroke identification sequence images to extract the local area; the local area is the area corresponding to the DWI high signal on the registered low-field magnetic resonance stroke identification sequence images; the DWI high signal is the image of the part with a value of 1 in the lesion segmentation result; the image intensity mean of the local area is calculated to obtain the local signal intensity mean; the intensity mean of the registered low-field magnetic resonance stroke identification sequence images in the entire brain is calculated to obtain the signal intensity mean in the whole brain; the local signal intensity mean is compared with the signal intensity mean in the whole brain to obtain the mismatch ratio; the mismatch ratio is used to provide a reference for doctors to assist doctors in determining the type of bleeding or ischemia of the target object.

[0010] Optionally, the low-field magnetic resonance stroke identification sequence image after skull stripping is registered to the DWI image after skull stripping to obtain the registered low-field magnetic resonance stroke identification sequence image, specifically including: using the ANTs registration method to keep the DWI image after skull stripping still, and transforming the low-field magnetic resonance stroke identification sequence image after skull stripping into the same physical space as the DWI image after skull stripping through Affine transformation to obtain the registered low-field magnetic resonance stroke identification sequence image; the Affine transformation includes rotation and translation operations.

[0011] Optionally, applying the lesion segmentation result to the registered low-field magnetic resonance stroke identification sequence image to extract the local area, specifically comprising: applying the lesion segmentation result to the registered low-field magnetic resonance stroke identification sequence image according to the formula Calculate and extract the local area; among them, represents the lesion segmentation result, , represents matrix element-wise multiplication, represents the registered low-field magnetic resonance stroke identification sequence image, , Represents a local area, , Indicates the dimension to which it belongs, and the size and shape of the image corresponding to each dimension, , , Respectively is a three-dimensional vector in real space, is a three-dimensional vector in real space, is a three-dimensional vector in real space, each dimension represents the height of the three-dimensional vector ,width and depth .

[0012] Optionally, calculating the image intensity mean of the local area to obtain the local signal intensity mean specifically includes: using the formula Calculate the image intensity mean of the local area to obtain the local signal intensity mean; wherein, represents the local signal intensity mean, Indicates the expectation. In three-dimensional space, through the coordinates ( , , ) uniquely identifies a point in space, represents the position along the height axis, Indicates the position along the width axis, represents the position along the depth axis, Indicates that it corresponds to ( , , ) coordinates, the intensity value of an image voxel.

[0013] Optionally, calculating the intensity mean of the registered low-field magnetic resonance stroke identification sequence image in the whole brain to obtain the signal intensity mean of the whole brain includes: using the formula Calculate the intensity mean of the registered low-field magnetic resonance stroke identification sequence image in the whole brain to obtain the signal intensity mean of the whole brain; wherein, represents the mean signal intensity in the whole brain, Representing images exist( , , ) is the intensity value of the voxel at .

[0014] Optionally, the local signal intensity mean is compared with the whole brain signal intensity mean to obtain a mismatch ratio, specifically comprising: using the formula Calculate the mismatch ratio; where, Represents the mismatch ratio; using The specific process of providing a reference for doctors and assisting them in determining the type of bleeding or ischemia of the target object includes: With high signal threshold For size comparison, if , it is determined that a high signal appears in the local area. If , it is determined that no high signal appears in the local area, combined with the situation of DWI high signal, if a high signal appears on the DWI image, but no high signal appears on the low-field magnetic resonance stroke identification sequence image, it is determined that the DWI image and the low-field magnetic resonance stroke identification sequence image do not match, and the hemorrhage or ischemia type of the target object is ischemia; if high signals appear on both the DWI image and the low-field magnetic resonance stroke identification sequence image, it is determined that the DWI image and the low-field magnetic resonance stroke identification sequence image match, and the hemorrhage or ischemia type of the target object is hemorrhage.

[0015] In a second aspect, the present application provides an MRI mismatch scoring device, which includes: an image acquisition module, used to acquire a DWI image and a low-field magnetic resonance stroke identification sequence image of a target object; a lesion segmentation module, used to apply a DWI lesion segmentation model to segment the DWI image to obtain a lesion segmentation result on the DWI image; the DWI lesion segmentation model is obtained by training a neural network based on a U-Net framework using a DWI acute ischemic lesion data set; the lesion segmentation result is a binary mask; an image skull stripping module, used to perform skull stripping on the DWI image and the low-field magnetic resonance stroke identification sequence image respectively, to obtain a DWI image after skull stripping and a low-field magnetic resonance stroke identification sequence image after skull stripping; an image registration module, used to register the low-field magnetic resonance stroke identification sequence image after skull stripping to the DWI image after skull stripping, Obtain the registered low-field magnetic resonance stroke identification sequence image; a local area extraction module is used to apply the lesion segmentation result to the registered low-field magnetic resonance stroke identification sequence image to extract the local area; the local area is the area corresponding to the DWI high signal on the registered low-field magnetic resonance stroke identification sequence image; the DWI high signal is the image of the part with a value of 1 in the lesion segmentation result; a local signal intensity mean calculation module is used to calculate the image intensity mean of the local area to obtain the local signal intensity mean; a whole-brain signal intensity mean calculation module is used to calculate the intensity mean of the registered low-field magnetic resonance stroke identification sequence image in the whole brain to obtain the whole-brain signal intensity mean; a mismatch ratio calculation module is used to compare the local signal intensity mean with the whole-brain signal intensity mean to obtain the mismatch ratio; the mismatch ratio is used to provide a reference for doctors to assist doctors in determining the type of bleeding or ischemia of the target object.

[0016] In a third aspect, the present application provides a computer device, 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 any one of the above-described MRI mismatch scoring methods.

[0017] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any one of the above-described MRI mismatch scoring methods.

[0018] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any one of the above-mentioned MRI mismatch scoring methods.

[0019] According to the specific embodiments provided in this application, this application has the following technical effects.

[0020] The present application provides an MRI mismatch scoring method, device, equipment, medium and product, which registers a low-field magnetic resonance stroke identification sequence image to a DWI image, applies the lesion segmentation result on the DWI image to the registered low-field magnetic resonance stroke identification sequence image, obtains an area corresponding to a DWI high signal on the registered low-field magnetic resonance stroke identification sequence image, calculates the image intensity mean of the area, compares the image intensity mean of the area with the intensity mean of the entire brain of the registered low-field magnetic resonance stroke identification sequence image, and obtains a mismatch ratio, thereby realizing DWI-low-field magnetic resonance stroke identification sequence mismatch identification based on DWI and a low-field magnetic resonance stroke identification sequence that can quickly and accurately judge hemorrhagic stroke. Compared with traditional imaging analysis and reliance on doctor's interpretation, the DWI-low-field magnetic resonance stroke identification sequence mismatch identification method can improve the efficiency of MRI mismatch scoring and the accuracy of MRI mismatch scoring results. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 A flowchart of an MRI mismatch scoring method provided in one embodiment of the present application.

[0023] Figure 2 This is the UNet segmentation network model diagram for this application.

[0024] Figure 3 Schematic diagram of the data processing flow of paired DWI-low-field MRI stroke identification sequence for this application.

[0025] Figure 4 This is a schematic diagram of the steps of the method for quickly identifying hemorrhage and ischemia using a DWI-low-field magnetic resonance stroke identification sequence based on a deep learning model in this application.

[0026] Figure 5 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] The purpose of this application is to provide an MRI mismatch scoring method, device, equipment, medium and product, which can perform DWI-low-field MRI stroke identification sequence mismatch identification based on DWI and low-field MRI stroke identification sequences to improve the efficiency of MRI mismatch scoring and the accuracy of MRI mismatch scoring results.

[0029] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0030] like Figure 1 As shown, the present application provides an MRI mismatch scoring method, including steps 101 to 108.

[0031] Step 101: Obtain DWI images and low-field magnetic resonance stroke identification sequence images of a target object.

[0032] Step 102: Apply the DWI lesion segmentation model to segment the DWI image to obtain a lesion segmentation result on the DWI image; the DWI lesion segmentation model is obtained by training a neural network based on a U-Net framework using a DWI acute ischemic lesion dataset; the lesion segmentation result is a binary mask.

[0033] Step 103: performing skull stripping on the DWI image and the low-field MRI stroke identification sequence image, respectively, to obtain the DWI image after skull stripping and the low-field MRI stroke identification sequence image after skull stripping.

[0034] Step 104: registering the low-field MRI stroke identification sequence images after skull stripping to the DWI images after skull stripping to obtain registered low-field MRI stroke identification sequence images.

[0035] The step 104 specifically includes: using the ANTs registration method to keep the DWI image after skull stripping still, and transforming the low-field MRI stroke identification sequence image after skull stripping into the same physical space as the DWI image after skull stripping through Affine transformation to obtain the registered low-field MRI stroke identification sequence image; Affine transformation includes rotation and translation operations.

[0036] Step 105: Apply the lesion segmentation result to the registered low-field MRI stroke identification sequence image to extract the local area; the local area is the area corresponding to the DWI high signal on the registered low-field MRI stroke identification sequence image; the DWI high signal is the image of the part with a value of 1 in the lesion segmentation result.

[0037] The step 105 specifically includes: comparing the lesion segmentation result with the registered low-field magnetic resonance stroke identification sequence image according to the formula Calculate and extract the local area; among them, represents the lesion segmentation result, , represents matrix element-wise multiplication, It represents the registered low-field magnetic resonance imaging sequence for stroke identification. , Represents a local area, , Indicates the dimension to which it belongs, and the size and shape of the image corresponding to each dimension, , , Respectively is a three-dimensional vector in real space, is a three-dimensional vector in real space, is a three-dimensional vector in real space, each dimension represents the height of the three-dimensional vector ,width and depth .

[0038] Step 106: Calculate the image intensity mean of the local area to obtain the local signal intensity mean.

[0039] The step 106 specifically includes: using the formula Calculate the image intensity mean of the local area to obtain the local signal intensity mean; wherein, represents the local signal intensity mean, Indicates the expectation. In three-dimensional space, through the coordinates ( , , ) uniquely identifies a point in space, represents the position along the height axis, Indicates the position along the width axis, represents the position along the depth axis, Indicates that it corresponds to ( , , ) coordinates, the intensity value of an image voxel.

[0040] Step 107: Calculate the intensity mean of the registered low-field magnetic resonance stroke identification sequence images throughout the brain to obtain the signal intensity mean of the entire brain.

[0041] The step 107 specifically includes: using the formula Calculate the intensity mean of the registered low-field MRI stroke identification sequence images throughout the brain to obtain the signal intensity mean of the whole brain; represents the mean signal intensity in the whole brain, Representing images exist( , , ) is the intensity value of the voxel at .

[0042] Step 108: Compare the local signal intensity mean with the whole brain signal intensity mean to obtain a mismatch ratio; the mismatch ratio is used to provide a reference for doctors to assist them in determining the type of hemorrhage or ischemia of the target object.

[0043] The step 108 specifically includes: using the formula Calculate the mismatch ratio; where, Represents the mismatch ratio; using The specific process of providing a reference for doctors and assisting them in determining the type of bleeding or ischemia of the target object includes: With high signal threshold For size comparison, if , it is judged that a high signal appears in the local area. , judge that there is no high signal in the local area, combined with the situation of DWI high signal, if high signal appears on the DWI image, but no high signal appears on the low-field magnetic resonance stroke identification sequence image, it is judged that the DWI image and the low-field magnetic resonance stroke identification sequence image do not match, and the hemorrhage or ischemia type of the target object is ischemia. If high signal appears on both the DWI image and the low-field magnetic resonance stroke identification sequence image, it is judged that the DWI image and the low-field magnetic resonance stroke identification sequence image match, and the hemorrhage or ischemia type of the target object is hemorrhage.

[0044] The technical solution of the present application is described below with a specific embodiment.

[0045] The MRI mismatch scoring method provided in the present application is a DWI-low-field magnetic resonance stroke identification sequence signal mismatch scoring algorithm based on a deep learning model (i.e., a DWI-low-field magnetic resonance stroke identification sequence mismatch identification algorithm based on a deep learning model). By applying the deep learning method to DWI-low-field magnetic resonance stroke identification sequence mismatch identification, the deep learning model is used to automatically extract DWI and low-field magnetic resonance stroke identification sequence image features, to achieve rapid analysis and evaluation, provide real-time diagnosis results, shorten the diagnosis time, overcome the limitation of traditional imaging that relies on the subjective judgment of doctors, establish an objective and quantifiable DWI-low-field magnetic resonance stroke identification sequence signal mismatch evaluation system, provide a reliable diagnostic basis, assist clinicians in making decisions, improve the accuracy and efficiency of the identification of hyperacute cerebral hemorrhage and cerebral infarction, reduce the risk of misdiagnosis and missed diagnosis, and provide a new technical means for the rapid and accurate identification of hyperacute cerebral hemorrhage and cerebral infarction.

[0046] This application constructs an indicator for calculating the high signal mismatch between diffusion weighted imaging (DWI) and low-field magnetic resonance stroke identification sequence, and provides a DWI-low-field magnetic resonance stroke identification sequence mismatch scoring model based on a deep learning model, which is used to quickly identify hemorrhage and ischemia in the hyperacute stage of stroke and assist doctors in diagnosis. This method can automatically segment the lesions displayed on the patient's DWI images and perform segmentation quickly; then, through the mismatch scoring module designed in this application, the score of the degree of mismatch between the lesion signal currently displayed on the DWI and the corresponding low-field magnetic resonance stroke identification sequence is compared and calculated.

[0047] This application uses the capabilities of the deep model to achieve DWI stroke lesion segmentation, and through the designed signal mismatch scoring algorithm, calculates a score used to describe the degree of mismatch to prompt doctors to identify the type of hemorrhage and ischemia in patients in the hyperacute stage. The process of the DWI-low-field magnetic resonance stroke identification sequence mismatch score method for rapid identification of hemorrhage and ischemia is briefly described as follows.

[0048] (1) Using the DWI acute ischemic lesion dataset, the Figure 2 The DWI-UNet segmentation network (i.e., a neural network based on the U-Net framework) shown in the figure is trained to obtain a DWI lesion segmentation model.

[0049] (2) Use the DWI lesion segmentation model to segment the DWI image and obtain the lesion segmentation results on the DWI.

[0050] (3) For a pair of DWI and low-field MRI stroke differential sequence images (the DWI images and low-field MRI stroke differential sequence images are corresponding), use the registration library in ANTs (Advanced Normalization Tools) to register the low-field MRI stroke differential sequence images to the DWI images. DWI sequences and low-field MRI stroke differential sequences can be quickly scanned within a few minutes using low-field mobile MRI equipment. Since multiple sequences are scanned for diagnosis when the same patient undergoes MRI scanning, the DWI and low-field MRI stroke differential sequences obtained from the same patient are corresponding.

[0051] (4) Apply the obtained segmentation result (binarized mask) to the registered low-field MRI stroke identification sequence image, calculate the mean image intensity of the part with a value of 1 in the mask, and then calculate the mean intensity of the entire brain of the low-field MRI stroke identification sequence, and compare the two means to obtain a ratio that can represent the signal difference inside and outside the region. Among them, the image intensity of the part with a value of 1 is obtained as follows: the segmentation mask is binary (only 0 and 1), and each pixel value of the image in the computer represents the voxel intensity value of the image at that location, so the segmentation mask is multiplied by the image (a matrix of the same shape and size), and the intensity value of the image corresponding to the segmentation mask of 0 becomes 0, and the part of 1 remains unchanged, that is, the part corresponding to the segmentation mask of 1 is taken out. For example, a method for calculating the mean intensity of the whole brain in a low-field MRI stroke identification sequence is used. For example, a low-field MRI stroke identification sequence image is represented by A, and its shape is a (13, 512, 512) matrix, and each element represents the image intensity value, for example, A[0, 0, 0] = 725. First, A is skull-stripped, such as Figure 3 As shown, we get an image of only the brain and calculate the overall mean: add up each element of the matrix and divide it by the total number of elements.

[0052] (5) According to the set threshold ( ), compare it with the ratio obtained in (4) and get the identification result.

[0053] The lesion segmentation model in this application is explained below.

[0054] The DWI-UNet lesion segmentation model (i.e., DWI lesion segmentation model) used in this application adopts improved network training based on UNet. First, a DWI acute ischemic lesion dataset is constructed, the original DWI images are selected, and abnormal data are removed. Then, the doctor manually outlines the lesion area as the lesion annotation of the corresponding image, which constitutes a lesion segmentation dataset with the original image data. The model constructed in this application is based on the traditional U-Net framework, and the lesion segmentation network is mainly composed of 4 downsampling and upsampling stages. Each downsampling layer includes a maximum pooling layer, two convolutional layers (convolution kernel size is 3, stride is 1, padding is 1), two InstanceNorm layers and two ReLU activation functions, and is constructed in the order of maximum pooling layer, convolution layer-1, InstanceNorm-1, ReLU-1, convolution layer-2, InstanceNorm-2, ReLU-2; each upsampling layer includes a deconvolution layer (convolution kernel size is 2, stride is 2, padding is 0), two convolutional layers (convolution kernel size is 3, stride is 1, padding is 1), two InstanceNorm layers and two ReLU activation functions, and is constructed in the order of deconvolution layer, convolution layer-1, InstanceNorm-1, ReLU-1, convolution layer-2, InstanceNorm-2, ReLU-2. In order to improve the network's ability to extract features for medical images, a dense connection method is also used in the downsampling stage, and a residual module is added at the bottleneck to ensure that the segmentation model can extract enough information for segmentation. Finally, the output of the network is passed through a Sigmoid activation function to limit the output range to [0, 1], and the value is considered to be the probability result of the lesion prediction. The reason for outputting the probability value when the model predicts is that the label is a binary mask of 0 or 1, 0 means that the place is normal, and 1 means that there is a lesion. When the model is trained, it is necessary to predict whether a certain place is a lesion. The output value is a number between 0 and 1. The larger the value, the higher the confidence of the model that there is a lesion, so the output of the model prediction is a probability value. The final output is to generate a segmentation mask based on a confidence threshold. For example, if it is set to 0.5, then when the model predicts that the probability of this place being a lesion is 0.6 (>0.5), it is considered to be a lesion, and the final output mask value of this place is set to 1, otherwise it is set to 0.

[0055] Weighted Binary Cross Entropy Loss and DICE loss are used during training. The total loss function of training is the sum of the two, and their calculation formulas are as follows.

[0056] .

[0057] .

[0058] in, represents the weighted cross entropy loss, represents the DICE loss, represents the weight parameter, Indicates the first pixels, Represents the true label value at the pixel. Represents the predicted value of the network at the pixel point, and N represents the total number of pixels in an image. Finally, the network is trained by reverse gradient propagation and gradient descent algorithms until the network converges. Using cross-validation, the average DICE value and average Hausdorff Distance value of the model segmentation results are evaluated on the test set, and the model with the best evaluation index, that is, the highest segmentation accuracy, is selected for parameter preservation, and finally the DWI-UNet lesion segmentation model is obtained.

[0059] The UNet lesion segmentation network is only one step in the scoring method. First, it is used to identify the lesion part in the DWI image, and then with the help of the lesion mask map, the corresponding calculation method is used to perform corresponding calculations on the low-field magnetic resonance stroke identification sequence images to obtain the corresponding indicators and achieve the final judgment.

[0060] The following is an explanation of the DWI-low-field magnetic resonance stroke identification sequence mismatch scoring method in this application.

[0061] For the DWI image and low-field magnetic resonance stroke identification sequence image of the same patient, it is considered that if a high signal is shown on DWI but not on the low-field magnetic resonance stroke identification sequence, it means a mismatch, and in this case the current patient can be directly identified as ischemic; conversely, if both DWI and the low-field magnetic resonance stroke identification sequence show high signals, they match, and in this case the current patient can be directly identified as hemorrhage. The traditional modality signal mismatch algorithm based on deep learning is usually calculated based on the segmentation results of different modalities, and it is very challenging to use the lesion segmentation method for the low-field magnetic resonance stroke identification sequence: first, the high signal area in the low-field magnetic resonance stroke identification sequence is not as significant as that in the DWI sequence, which will increase the difficulty of model learning and cannot guarantee the accuracy of segmentation; secondly, if the model size and complexity are increased to enhance its learning ability, it will lead to a substantial increase in training resources and reasoning time, and lose the original goal of this application to quickly and economically identify ischemia; finally, the construction of the low-field magnetic resonance stroke identification sequence lesion data set will be more time-consuming and labor-intensive than other modalities.

[0062] Therefore, the present application constructs an indicator based on the DWI high-signal lesion area to infer whether the corresponding low-field magnetic resonance stroke identification sequence area has a high signal, and then combines the DWI signal situation (the present application is aimed at the disease, DWI will definitely show a high signal, and the absence of a high signal in DWI indicates that the patient is healthy and does not need to be identified for hemorrhage and ischemia) for comprehensive evaluation to achieve rapid hemorrhage and ischemia identification. The specific indicator calculation method is described below: Usually, the images of the DWI and low-field magnetic resonance stroke identification sequence of the same patient are not completely aligned, that is, for the same coordinate, the two may correspond to different positions. Therefore, it is necessary to align the same anatomical structures in the two images, and when expressed in coordinates, it is hoped that they are on the same coordinates. For example Figure 3 As shown in the figure, the Synthstrip model is first used to perform skull stripping on the DWI and low-field MRI stroke identification sequences respectively. This is not only conducive to the accuracy of subsequent registration, but also a necessary step for calculating the high signal intensity mean ratio on the low-field MRI stroke identification sequence, because the skull part usually has a higher intensity signal than other brain tissues; then the registration method in the ANTs library is used to register the low-field MRI stroke identification sequence image to the DWI image using Affine transformation to obtain the registered low-field MRI stroke identification sequence image. The image shape is recorded as ( , , ), respectively represent the height, width, and number of slices (i.e., depth). Among them, the low-field MRI stroke identification sequence is registered to DWI, which means that DWI remains unchanged, and the low-field MRI stroke identification sequence is transformed into the same physical space as DWI through operations such as rotation and translation. The new image obtained is the registered low-field MRI stroke identification sequence image.

[0063] At this point, the lesion mask obtained from DWI segmentation Comparison of registered low-field MRI stroke identification sequences Calculate according to the following formula and extract the low-field MRI stroke identification sequence area corresponding to the DWI high signal .

[0064] .

[0065] in, Indicates the dimensions to which it belongs, and the size and shape of the image corresponding to each dimension. For example express: It is a three-dimensional vector in real space, each dimension represents its height, width and depth. DWI high signal refers to the image with a value of 1 in the lesion segmentation result on DWI.

[0066] Calculate the local signal strength mean and the mean signal intensity of the whole brain .

[0067] .

[0068] .

[0069] in, To express expectation, such as As mentioned above, in three-dimensional space, through coordinates ( , , ) uniquely identifies a point in space, represents the position along the height axis, Indicates the position along the width axis, represents the position along the depth axis, Indicates that it corresponds to ( , , ) coordinates, the intensity value of an image voxel. For 2D images, ( , ) can determine a position, and the point corresponding to the image position is usually called a pixel point; for 3D images, ( , , ) can determine a position, and the point corresponding to the image position is usually called a voxel point. The local signal intensity mean refers to the local mean of the low-field MRI stroke identification sequence after registration corresponding to the DWI value of 1, and the whole brain signal intensity mean refers to the intensity mean of the whole brain of the low-field MRI stroke identification sequence.

[0070] Calculate the ratio of local strength to global strength .

[0071] .

[0072] By and a high signal threshold that can be set based on the image data For comparison, if If the intensity signal value of the local area is significantly higher than the average intensity value of the entire brain, it can be considered that a high signal appears in this area; on the contrary, there is no high signal in the local area. Therefore, combined with the high signal of DWI, the identification of ischemia can be achieved. Based on the characteristics of the images, experienced medical experts suggested that adjustments can be made based on the data.

[0073] like Figure 4As shown, the present application proposes a method for quickly identifying hemorrhage and ischemia using a DWI-low-field magnetic resonance stroke identification sequence based on a deep learning model, and the steps are as follows.

[0074] S1: Training lesion segmentation model.

[0075] The constructed segmentation dataset is used to train the DWI-UNet lesion segmentation network to obtain an accurate lesion segmentation model.

[0076] S2: Perform lesion segmentation on DWI images.

[0077] Use the trained DWI-UNet segmentation network to perform lesion segmentation on the DWI image (DWI image without skull removal). The input data accepted by the trained UNet segmentation network is the DWI image without skull removal. When using the trained model to segment the DWI image, there is no need to perform additional operations on the DWI. You only need to send it to the UNet segmentation network to get the segmentation result.

[0078] S3: Preprocess low-field MRI stroke identification sequence images and DWI images.

[0079] The skull was stripped using Synthstrip, and the ANTs registration method was used to register the low-field MRI stroke identification sequence with DWI using Affine transformation to obtain the registered low-field MRI stroke identification sequence brain image and the corresponding DWI brain image. The preprocessing of the low-field MRI stroke identification sequence and DWI is for the subsequent calculation steps.

[0080] S4: Calculation of signal intensity in low-field MRI stroke identification sequence.

[0081] The lesion segmentation mask obtained in step S2 is calculated on the low-field magnetic resonance stroke identification sequence images according to the image signal intensity mean ratio calculation formula constructed above to obtain .

[0082] S5: Tips for differentiating between bleeding and ischemia.

[0083] Will With the high signal threshold set The size comparison is combined with the DWI high signal situation to obtain the final hemorrhage or ischemia type judgment result. The DWI-low field magnetic resonance imaging stroke identification sequence signal mismatch scoring algorithm based on the deep learning model in this application finally obtains a mismatch ratio, and the ischemia judgment is realized according to the size relationship between the ratio and the threshold, which is used to assist the doctor's judgment.

[0084] This application constructs an indicator of lesion visibility mismatch between DWI and low-field MRI stroke differentiation sequences, and develops a DWI-low-field MRI stroke differentiation sequence mismatch identification algorithm based on a deep learning model to quickly identify hyperacute cerebral hemorrhage and cerebral infarction to assist clinicians in making decisions.

[0085] The key points of this application are: 1. This application constructs a model based on a deep learning model for DWI-low-field magnetic resonance stroke identification sequence signal mismatch identification of hemorrhage and ischemia. The model has automated segmentation and signal mismatch identification capabilities, which can help doctors quickly identify the type of hemorrhage and ischemia in patients in the hyperacute stage and assist in treatment decisions. 2. This application constructs a method based on DWI high-signal area identification corresponding to the high signal of the low-field magnetic resonance stroke identification sequence area, which is used to quantify the degree of matching of the high signal of the lesion displayed on the DWI image and the low-field magnetic resonance stroke identification sequence image.

[0086] The advantages of this application are: 1. The DWI-low-field magnetic resonance stroke identification sequence mismatch identification model based on the deep learning model of this application helps doctors quickly identify the type of hemorrhage and ischemia in patients in the hyperacute period, realizes the automated analysis of medical imaging data without manual intervention, overcomes the limitation of traditional reliance on doctors' subjective judgment, shortens the diagnosis time, and helps reduce the risk of misdiagnosis and missed diagnosis. 2. The mismatch identification model of this application is based on deep learning, which can accurately identify the phenomenon of high signal mismatch between DWI and low-field magnetic resonance stroke identification sequence images, establishes an objective mismatch evaluation system, provides a reliable basis for diagnosis, and assists doctors to better formulate treatment strategies according to the identification type. 3. With the help of this application, the accuracy and efficiency of the identification of hyperacute cerebral hemorrhage and cerebral infarction can be improved, and doctors can be further assisted to quickly make treatment decisions, which is expected to have a positive clinical impact on the treatment results of patients, increase the opportunity for patients to obtain timely intervention, and help improve the survival rate and recovery rate of patients.

[0087] Based on the same inventive concept, the embodiment of the present application also provides an MRI mismatch scoring device for implementing the MRI mismatch scoring method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more MRI mismatch scoring device embodiments provided below can refer to the limitations of the MRI mismatch scoring method above, and will not be repeated here.

[0088] In an exemplary embodiment, an MRI mismatch scoring device is provided, comprising the following modules.

[0089] The image acquisition module is used to acquire DWI images and low-field magnetic resonance stroke identification sequence images of the target object.

[0090] The lesion segmentation module is used to apply the DWI lesion segmentation model to segment the DWI image and obtain the lesion segmentation result on the DWI image; the DWI lesion segmentation model is obtained by training a neural network based on the U-Net framework using the DWI acute ischemic lesion data set; the lesion segmentation result is a binary mask.

[0091] The image skull stripping module is used to perform skull stripping on the DWI image and the low-field magnetic resonance stroke identification sequence image respectively, so as to obtain the DWI image after skull stripping and the low-field magnetic resonance stroke identification sequence image after skull stripping.

[0092] The image registration module is used to register the low-field magnetic resonance stroke identification sequence images after skull stripping to the DWI images after skull stripping to obtain the registered low-field magnetic resonance stroke identification sequence images.

[0093] The local area extraction module is used to apply the lesion segmentation results to the registered low-field MRI stroke identification sequence images to extract the local area; the local area is the area corresponding to the DWI high signal on the registered low-field MRI stroke identification sequence images; the DWI high signal is the image with a value of 1 in the lesion segmentation result.

[0094] The local signal intensity mean calculation module is used to calculate the image intensity mean of the local area to obtain the local signal intensity mean.

[0095] The whole-brain signal intensity mean calculation module is used to calculate the intensity mean of the whole brain of the registered low-field magnetic resonance stroke identification sequence images to obtain the whole-brain signal intensity mean.

[0096] The mismatch ratio calculation module is used to compare the local signal intensity mean with the whole brain signal intensity mean to obtain the mismatch ratio; the mismatch ratio is used to provide a reference for doctors to assist them in determining the type of bleeding or ischemia of the target object.

[0097] Among them, the image registration module specifically includes: an image registration unit, which is used to use the ANTs registration method to keep the DWI image after skull stripping still, and transform the low-field magnetic resonance stroke identification sequence image after skull stripping into the same physical space as the DWI image after skull stripping through Affine transformation to obtain the registered low-field magnetic resonance stroke identification sequence image; Affine transformation includes rotation and translation operations.

[0098] The local area extraction module specifically includes: a local area extraction unit, which is used to combine the lesion segmentation result with the registered low-field magnetic resonance stroke identification sequence image according to the formula Calculate and extract the local area; among them, represents the lesion segmentation result, , represents matrix element-wise multiplication, It represents the registered low-field magnetic resonance imaging sequence for stroke identification. , Represents a local area, , Indicates the dimension to which it belongs, and the size and shape of the image corresponding to each dimension, , , Respectively is a three-dimensional vector in real space, is a three-dimensional vector in real space, is a three-dimensional vector in real space, each dimension represents the height of the three-dimensional vector ,width and depth .

[0099] The local signal strength mean value calculation module specifically includes: a local signal strength mean value calculation unit, which is used to use the formula Calculate the image intensity mean of the local area to obtain the local signal intensity mean; wherein, represents the local signal intensity mean, Indicates the expectation. In three-dimensional space, through the coordinates ( , , ) uniquely identifies a point in space, represents the position along the height axis, Indicates the position along the width axis, represents the position along the depth axis, Indicates that it corresponds to ( , , ) coordinates, the intensity value of an image voxel.

[0100] The whole brain signal strength mean calculation module specifically includes: the whole brain signal strength mean calculation unit, which is used to use the formula Calculate the intensity mean of the registered low-field MRI stroke identification sequence images throughout the brain to obtain the signal intensity mean of the whole brain; represents the mean signal intensity in the whole brain, Representing images exist( , , ) is the intensity value of the voxel at .

[0101] The mismatch ratio calculation module specifically includes: a mismatch ratio calculation unit, which is used to use the formula Calculate the mismatch ratio; where, Represents the mismatch ratio; using The specific process of providing a reference for doctors and assisting them in determining the type of bleeding or ischemia of the target object includes: With high signal threshold For size comparison, if , it is judged that a high signal appears in the local area. , judge that there is no high signal in the local area, combined with the situation of DWI high signal, if high signal appears on the DWI image, but no high signal appears on the low-field magnetic resonance stroke identification sequence image, it is judged that the DWI image and the low-field magnetic resonance stroke identification sequence image do not match, and the hemorrhage or ischemia type of the target object is ischemia. If high signal appears on both the DWI image and the low-field magnetic resonance stroke identification sequence image, it is judged that the DWI image and the low-field magnetic resonance stroke identification sequence image match, and the hemorrhage or ischemia type of the target object is hemorrhage.

[0102] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Wherein, the processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. Wherein, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store MRI mismatch scoring data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a MRI mismatch scoring method is implemented.

[0103] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0104] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0105] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0106] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0107] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0108] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0109] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0110] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for scoring MRI mismatch, characterized in that: The MRI mismatch scoring method includes: Acquire DWI images and low-field magnetic resonance stroke identification sequence images of the target object; The DWI image is segmented by applying a DWI lesion segmentation model to obtain a lesion segmentation result on the DWI image; the DWI lesion segmentation model is obtained by training a neural network based on a U-Net framework using a DWI acute ischemic lesion data set; the lesion segmentation result is a binary mask; Performing skull stripping on the DWI image and the low-field magnetic resonance stroke identification sequence image respectively to obtain the DWI image after skull stripping and the low-field magnetic resonance stroke identification sequence image after skull stripping; Registering the low-field magnetic resonance stroke identification sequence images after skull stripping to the DWI images after skull stripping to obtain registered low-field magnetic resonance stroke identification sequence images; Apply the lesion segmentation result to the registered low-field magnetic resonance stroke identification sequence image to extract a local area; the local area is an area corresponding to a DWI high signal on the registered low-field magnetic resonance stroke identification sequence image; the DWI high signal is an image of a portion of which the value is 1 in the lesion segmentation result; Calculating the image intensity mean of the local area to obtain the local signal intensity mean; Calculating the intensity mean of the registered low-field magnetic resonance stroke identification sequence images throughout the brain to obtain the signal intensity mean of the whole brain; The local signal intensity mean is compared with the whole brain signal intensity mean to obtain a mismatch ratio; the mismatch ratio is used to provide a reference for doctors to assist them in determining the type of hemorrhage or ischemia of the target object.

2. The MRI mismatch scoring method according to claim 1, characterized in that: The low-field magnetic resonance stroke identification sequence image after skull stripping is registered with the DWI image after skull stripping to obtain the registered low-field magnetic resonance stroke identification sequence image, specifically comprising: The ANTs registration method is used to keep the DWI image after skull stripping still, and the low-field magnetic resonance stroke identification sequence image after skull stripping is transformed into the same physical space as the DWI image after skull stripping through Affine transformation to obtain the registered low-field magnetic resonance stroke identification sequence image; the Affine transformation includes rotation and translation operations.

3. The MRI mismatch scoring method according to claim 1, characterized in that: Applying the lesion segmentation result to the registered low-field magnetic resonance stroke identification sequence image to extract the local area specifically includes: The lesion segmentation result and the registered low-field magnetic resonance stroke identification sequence image are combined according to the formula Calculate and extract the local area; among them, represents the lesion segmentation result, , represents matrix element-wise multiplication, represents the registered low-field magnetic resonance stroke identification sequence image, , Represents a local area, , Indicates the dimension to which it belongs, and the size and shape of the image corresponding to each dimension, , , Respectively is a three-dimensional vector in real space, is a three-dimensional vector in real space, is a three-dimensional vector in real space, each dimension represents the height of the three-dimensional vector ,width and depth .

4. The MRI mismatch scoring method according to claim 3, characterized in that: Calculating the image intensity mean of the local area to obtain the local signal intensity mean, specifically including: Using the formula Calculate the image intensity mean of the local area to obtain the local signal intensity mean; wherein, represents the local signal intensity mean, Indicates the expectation. In three-dimensional space, through the coordinates ( , , ) uniquely identifies a point in space, represents the position along the height axis, Indicates the position along the width axis, represents the position along the depth axis, Indicates that it corresponds to ( , , ) coordinates, the intensity value of an image voxel.

5. The MRI mismatch scoring method according to claim 4, characterized in that: Calculating the intensity mean of the registered low-field magnetic resonance stroke identification sequence image in the whole brain to obtain the signal intensity mean of the whole brain, specifically including: Using the formula Calculate the intensity mean of the registered low-field magnetic resonance stroke identification sequence image in the whole brain to obtain the signal intensity mean of the whole brain; wherein, represents the mean signal intensity in the whole brain, Representing images exist( , , ) is the intensity value of the voxel at .

6. The MRI mismatch scoring method according to claim 5, characterized in that: Comparing the local signal intensity mean with the whole brain signal intensity mean to obtain a mismatch ratio, specifically including: Using the formula Calculate the mismatch ratio; where, Represents the mismatch ratio; using The specific process of providing a reference for doctors and assisting them in determining the type of bleeding or ischemia of the target object includes: With high signal threshold For size comparison, if , it is determined that a high signal appears in the local area. If , it is determined that no high signal appears in the local area, combined with the situation of DWI high signal, if a high signal appears on the DWI image, but no high signal appears on the low-field magnetic resonance stroke identification sequence image, it is determined that the DWI image and the low-field magnetic resonance stroke identification sequence image do not match, and the hemorrhage or ischemia type of the target object is ischemia; if high signals appear on both the DWI image and the low-field magnetic resonance stroke identification sequence image, it is determined that the DWI image and the low-field magnetic resonance stroke identification sequence image match, and the hemorrhage or ischemia type of the target object is hemorrhage.

7. An MRI mismatch scoring device, characterized in that: The MRI mismatch scoring device comprises: An image acquisition module, used to acquire DWI images and low-field magnetic resonance stroke identification sequence images of the target object; A lesion segmentation module is used to segment the DWI image using a DWI lesion segmentation model to obtain a lesion segmentation result on the DWI image; the DWI lesion segmentation model is obtained by training a neural network based on a U-Net framework using a DWI acute ischemic lesion data set; the lesion segmentation result is a binary mask; An image skull stripping module is used to perform skull stripping on the DWI image and the low-field magnetic resonance stroke identification sequence image, respectively, to obtain the DWI image after skull stripping and the low-field magnetic resonance stroke identification sequence image after skull stripping; An image registration module, used for registering the low-field magnetic resonance stroke identification sequence image after skull stripping to the DWI image after skull stripping to obtain the registered low-field magnetic resonance stroke identification sequence image; A local area extraction module is used to apply the lesion segmentation result to the registered low-field magnetic resonance stroke identification sequence image to extract a local area; the local area is the area corresponding to the DWI high signal on the registered low-field magnetic resonance stroke identification sequence image; the DWI high signal is the image of the part with a value of 1 in the lesion segmentation result; A local signal intensity mean value calculation module is used to calculate the image intensity mean value of the local area to obtain the local signal intensity mean value; A whole-brain signal intensity mean calculation module is used to calculate the intensity mean of the whole brain of the registered low-field magnetic resonance stroke identification sequence image to obtain the whole-brain signal intensity mean; The mismatch ratio calculation module is used to compare the local signal intensity mean with the whole brain signal intensity mean to obtain a mismatch ratio; the mismatch ratio is used to provide a reference for doctors to assist doctors in determining the type of bleeding or ischemia of the target object.

8. A computer device 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 MRI mismatch scoring method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the MRI mismatch scoring method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the MRI mismatch scoring method according to any one of claims 1 to 6 is implemented.

Images