Automatic detection method and device for cerebral cortex micro-infarction

Abstract

The invention provides an automatic detection method and device for cerebral cortex micro-infarction (CMIs). According to the method and the device, the CMIs of a patient are intelligently identified and judged by using an artificial intelligence (AI) model in combination with a clinical diagnosis standard of the CMIs, and the method and the device aim at eliminating errors and low efficiency possibly caused by subjective judgment of a doctor, so that rapid, objective and accurate evaluation of the CMIs is realized. According to the method, on the basis of a magnetic resonance (MRI) image, cerebral cortex signal characteristics of a patient under a T1 sequence, a liquid attenuation inversion recovery (FLAIR) sequence and a double inversion recovery (DIR) sequence are intelligently analyzed through an AI model, rapid clinical diagnosis support can be provided for doctors, dependence on professional skills and clinical experience is reduced, the diagnosis efficiency of CMIs is remarkably improved, and the diagnosis accuracy of the CMIs is improved. The related research progress and clinical application are accelerated.

Description

Technical Field

[0001] The present invention relates to the technical field of disease detection, and in particular to an automatic detection method for cerebral cortical microinfarction, an automatic detection device and detection equipment for cerebral cortical microinfarction, and a computer-readable storage medium. Background Art

[0002] Cortical cerebral microinfarcts (CMIs) are the most common type of cerebral infarction, which is more common in the elderly and is independently associated with cognitive impairment and dementia. The incidence of cerebral cortical microinfarcts in patients with vascular dementia (VaD), Alzheimer's disease (AD) and elderly people without dementia is 62%, 43% and 24%, respectively.

[0003] However, due to the limitations of previous in vivo detection methods, in vivo research on CMIs is still in its initial stage, its etiology and disability mechanism have not yet been elucidated, and there is currently a lack of effective prevention and treatment methods.

[0005] Therefore, it is crucial to clarify the causes and disability mechanisms of CMIs and take targeted prevention and treatment measures to reduce their incidence and disability rates, which is also the current challenge.

[0006] The clinical diagnostic criteria for CMIs are:

[0007] 1. Location: limited to the cerebral cortex;

[0008] 2. Size: 0.5-4mm;

[0009] 3. As attached Figure 1 The signal characteristics shown are: low signal in T1 sequence, high signal in FLAIR sequence, and high signal in DIR sequence.

[0010] In clinical practice, the clinical diagnosis method of CMIs mainly relies on doctors to manually identify and mark the location of CMIs lesions in the patient's MRI images. This method relies on a high level of professionalism and clinical experience, resulting in certain differences in the diagnostic results of CMIs. In addition, doctors need to identify each MRI image with the naked eye, which is time-consuming and labor-intensive. The above pain points significantly limit the speed of standardized diagnosis and treatment of CMIs and high-quality research on large samples, making the research on CMIs progress slowly. To date, there is still a lack of effective prevention and treatment methods, and the mortality and disability rates remain high. Summary of the invention

[0011] In order to solve the technical problems existing in the prior art, the embodiments of the present invention provide the following technical solutions:

[0012] In one aspect, a method for automatically detecting cerebral cortical microinfarction is provided, the method being implemented by a detection device, the method comprising:

[0013] S1, obtaining head MRI images under MRI scanning sequences: T1, FLAIR and DIR sequences, and using the head MRI images as input to build a corresponding CMIs recognition model based on CNN model training;

[0014] S2, collecting cerebral cortex cranial MRI images of clinical patients under MRI scanning sequences: T1, FLAIR and DIR, and inputting them into the CMIs recognition model, and using the CMIs recognition model to recognize and output cerebral cortex MRI scanning signal features under T1, FLAIR and DIR respectively;

[0015] S3. Compare the cerebral cortex MRI scan signal characteristics of clinical patients with the CMIs signal characteristics of clinical diagnostic standards to determine whether the clinical patients meet the CMIs criteria. The clinical diagnostic criteria are as follows: (1) Location: limited to the cerebral cortex; (2) Size: 0.5-4 mm; (3) Signal: T1 low signal, FLAIR high signal, DIR high signal;

[0016] S4. If the cerebral cortex MRI scanning signal characteristics under three different scanning sequences appear on the cerebral cortex head MRI image of the clinical patient, the CMIs lesions at each contour position are located and marked to obtain the corresponding CMIs lesion areas: C1, C2 and C3;

[0017] S5. Based on the spatial registration method, the CMIs lesion areas: C1, C2 and C3 are registered to determine the CMIs lesions of the clinical patient and mark them on the cerebral cortex head MRI image of the clinical patient;

[0018] S6. Output the cerebral cortex cranial MRI image with the CMIs lesions marked.

[0019] As a preferred embodiment of the present invention, in step S3, when determining whether the clinical patient meets the CMIs requirement, the following steps are further included:

[0020] If not, further determining the number of CMIs signal features of the clinical diagnostic standard matched by the cerebral cortex MRI scan signal features;

[0021] If none of the cerebral cortex MRI scan signal features matches the CMIs signal features of the clinical diagnostic standard, directly outputting the cerebral cortex head MRI image;

[0022] If at least one of the cerebral cortex MRI scan signal features matches the CMIs signal feature of the clinical diagnostic standard, the contour position of the matched cerebral cortex MRI scan signal feature is marked on the cerebral cortex head MRI image of the clinical patient, and the cerebral cortex head MRI image with the marked contour position is output.

[0023] As a preferred embodiment of the present invention, in step S5, after registering the Cortical CMI lesion areas: C1, C2 and C3 based on the spatial registration method to determine the CMIs lesions of clinical patients, it also includes:

[0024] Acquire a registered cranial MRI image of the CMIs lesion of a clinical patient;

[0025] Based on the three-dimensional reconstruction technology, the cranial MRI image of the CMIs lesion is three-dimensionally reconstructed to construct a three-dimensional CMIs lesion model corresponding to the CMIs lesion;

[0026] Output the three-dimensional CMIs lesion model of clinical patients.

[0027] As a preferred embodiment of the present invention, in step S5, after registering the CMIs lesion areas: C1, C2 and C3 based on the spatial registration method to determine the CMIs lesions of clinical patients, it also includes:

[0028] Saving the three-dimensional CMIs lesion model in OBJ format;

[0029] Importing the three-dimensional CMIs lesion model into MeshLab, and using the pyvista library in MeshLab to process and calculate the volume of the three-dimensional CMIs lesion model;

[0030] The volume of the three-dimensional CMIs lesion model is marked on the cerebral cortex cranial MRI image of a clinical patient.

[0031] As a preferred embodiment of the present invention, in step S1, respectively obtaining the patient's T1, FLAIR and DIR cranial MRI images, and using the cranial MRI images as input to construct a corresponding CMIs recognition model based on CNN model training, comprises:

[0032] The cranial MRI images of several CMIs patients were collected under the MRI scanning sequences: T1, FLAIR and DIR and preprocessed;

[0033] Performing feature engineering on the head MRI images under T1, FLAIR and DIR scanning sequences, respectively extracting the cerebral cortex MRI scanning signal features on the head MRI images under T1, FLAIR and DIR scanning sequences;

[0034] The extracted cerebral cortex MRI scanning signal features are used as input to a preset CNN model, and the cerebral cortex MRI scanning signal features are trained and learned using a convolutional neural network to obtain the CMIs recognition model;

[0035] The CMIs recognition model is verified using the preset cerebral cortex head MRI images under the MRI scanning sequences: T1, FLAIR and DIR;

[0036] If the verification is qualified, the recognition model is deployed on the background server.

[0037] As a preferred embodiment of the present invention, the S3, comparing the cerebral cortex MRI scan signal characteristics of the clinical patient with the CMIs signal characteristics of the clinical diagnostic standard to determine whether the clinical patient meets the CMIs, comprises:

[0038] Read the cerebral cortex MRI scanning signal characteristics of clinical patients under T1, FLAIR and DIR, and analyze and obtain the position, size and signal characteristics of the cerebral cortex MRI scanning signal characteristics;

[0039] The cerebral cortex MRI scan signal features under T1, FLAIR and DIR are compared with the position, size and signal features of the CMIs signal features of the preset clinical diagnostic standard one by one to determine whether they are consistent:

[0040] If they are consistent, the clinical patient is judged to meet CMIs;

[0041] Otherwise it does not meet the requirements.

[0042] As a preferred embodiment of the present invention, S4, if CMIs is met, the CMIs lesions at each contour position are located and marked according to the contour position of the cerebral cortex MRI scan signal characteristics of the clinical patient, and the corresponding CMIs lesion areas: C1, C2 and C3 are obtained, including:

[0043] According to the characteristic images shown by the cerebral cortex MRI scanning signal characteristics under T1, FLAIR and DIR, respectively, the image contours displayed by each of the cerebral cortex MRI scanning signal characteristics on the cerebral cortex head MRI image are obtained;

[0044] Based on the same positioning reference, the contours of each image were positioned and marked to obtain the CMIs lesion areas of the cerebral cortex under T1, FLAIR and DIR: C1, C2 and C3.

[0045] As a preferred embodiment of the present invention, S5, based on the spatial registration method, the CMIs lesion areas: C1, C2 and C3 are registered, the CMIs lesions of clinical patients are determined and marked on the cerebral cortex head MRI image of clinical patients, including:

[0046] The center point of the MRI scanning system is used as a reference;

[0047] Determine respectively the first spatial relationship, the second spatial relationship and the third spatial relationship between the CMIs lesion areas: C1, C2 and C3 and the benchmark;

[0048] The first spatial relationship, the second spatial relationship and the third spatial relationship are integrated to obtain the spatial coordinate relationship of the CMIs lesion of the clinical patient;

[0049] According to the spatial coordinate relationship, the CMIs lesion of the clinical patient is positioned and marked on the cerebral cortex head MRI image of the clinical patient.

[0050] On the other hand, a device for automatically detecting cerebral cortical microinfarction is provided, and the device for automatically detecting cerebral cortical microinfarction is used to implement the above-mentioned method for automatically detecting cerebral cortical microinfarction, and the device comprises:

[0051] MRI system, used to collect cerebral cortex cranial MRI images of clinical patients under MRI scanning sequences: T1, FLAIR and DIR, and input them into the backend server;

[0052] Backend server, including:

[0053] HIS system, used to record and save the cerebral cortex head MRI images of clinical patients;

[0054] The Cortical CMI recognition module is used to use the preset CMIs recognition model to identify and output the cerebral cortex MRI scanning signal characteristics of clinical patients under T1, FLAIR and DIR;

[0055] A feature comparison module is used to compare the cerebral cortical MRI scan signal features of clinical patients with the CMIs signal features of clinical diagnostic standards to determine whether the clinical patients meet the Cortical CMI;

[0056] A lesion marking module is used to locate and mark the CMIs lesions at each contour position according to the contour position of the cerebral cortex MRI scan signal characteristics of the clinical patient if CMIs is met, and obtain the corresponding CMIs lesion areas: C1, C2 and C3;

[0057] A lesion registration module is used to register the CMIs lesion areas: C1, C2 and C3 based on a spatial registration method, determine the CMIs lesions of clinical patients and mark them on the cerebral cortex head MRI image of the clinical patients;

[0058] The output module is used to output the cerebral cortex head MRI image with the CMIs lesions marked.

[0059] On the other hand, a detection device is provided, comprising: a processor; a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any one of the above-mentioned automatic detection methods for cerebral cortical microinfarction is implemented.

[0060] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, and the at least one instruction is loaded and executed by a processor to implement any one of the above-mentioned automatic detection methods for cerebral cortical microinfarction.

[0061] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0062] Based on the implementation of the present invention, the present invention uses the CMIs recognition model to identify and output the cerebral cortex MRI scanning signal features of the cerebral cortex under T1, FLAIR and DIR respectively; compare the cerebral cortex MRI scanning signal features of clinical patients with the CMIs signal features of clinical diagnostic standards to determine whether the clinical patients meet the CMIs; if they meet the CMIs, then according to the contour position of the cerebral cortex MRI scanning signal features of the clinical patients, the CMIs lesions at each contour position are located and marked to obtain the corresponding CMIs lesion areas: C1, C2 and C3; based on the spatial registration method, the CMIs lesion areas: C1, C2 and C3 are registered to determine the CMIs lesions of clinical patients and mark them on the cerebral cortex head MRI image of the clinical patients; output the cerebral cortex head MRI image with the CMIs lesions marked. Therefore, combined with the clinical diagnostic standards of CMIs, the AI ​​model is used to intelligently identify and judge the CMIs of patients, so as to make a clinically objective CMIs diagnosis and avoid being affected by the errors of the subjective judgment of doctors. This solution uses an AI model to intelligently identify the patient's cerebral cortex MRI scan signal characteristics under T1, FLAIR and DIR, which can provide doctors with rapid clinical diagnosis, reduce the requirements for expertise and experience, improve the diagnostic identification efficiency of CMIs, and speed up the research and diagnosis of CMIs. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0064] Figure 1 It is a schematic diagram of the cerebral cortex MRI scan signal characteristics under T1, FLAIR and DIR (clinical diagnostic criteria for CMIs);

[0065] Figure 2 is a flow chart of an automatic detection method for cerebral cortical microinfarction provided by an embodiment of the present invention;

[0066] Figure 3 This is a block diagram of an automatic detection device for cerebral cortical microinfarction provided by an embodiment of the present invention;

[0067] Figure 4 It is a flowchart of a training model provided by an embodiment of the present invention;

[0068] Figure 5It is a schematic diagram of marking a CMIs lesion area of ​​a cerebral cortex under T1, FLAIR and DIR respectively provided by an embodiment of the present invention;

[0069] Figure 6 It is a structural schematic diagram of a detection device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0070] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0071] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0072] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0073] In the embodiments of the present invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are consistent.

[0074] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0075] The embodiment of the present invention provides an automatic detection method for cerebral cortical microinfarction, which can be implemented by a detection device, and the detection device can be a terminal or a server.

[0076] like Figure 2 The flowchart of the automatic detection method for cerebral cortical microinfarction shown in the figure may include the following steps:

[0077] In one aspect, a method for automatically detecting cerebral cortical microinfarction is provided, the method being implemented by a detection device, the method comprising:

[0078] S1, respectively obtain the patient's T1, FLAIR and DIR cranial MRI images, and use the cranial MRI images as input to build a corresponding CMIs recognition model based on CNN model training;

[0079] S2, collecting cerebral cortex cranial MRI images of clinical patients under MRI scanning sequences: T1, FLAIR and DIR, and inputting them into the CMIs recognition model, and using the CMIs recognition model to recognize and output cerebral cortex MRI scanning signal features under T1, FLAIR and DIR respectively;

[0080] S3. Compare the cerebral cortex MRI scan signal characteristics of clinical patients with the CMIs signal characteristics of clinical diagnostic standards to determine whether the clinical patients meet the CMIs criteria. The clinical diagnostic criteria are as follows: (1) Location: limited to the cerebral cortex; (2) Size: 0.5-4 mm; (3) Signal: T1 low signal, FLAIR high signal, DIR high signal;

[0081] S4. If the cerebral cortex MRI scanning signal characteristics under three different scanning sequences appear on the cerebral cortex head MRI image of the clinical patient, the CMIs lesions at each contour position are located and marked to obtain the corresponding CMIs lesion areas: C1, C2 and C3;

[0082] S5. Based on the spatial registration method, the CMIs lesion areas: C1, C2 and C3 are registered to determine the CMIs lesions of the clinical patient and mark them on the cerebral cortex head MRI image of the clinical patient;

[0083] S6. Output the cerebral cortex cranial MRI image with the CMIs lesions marked.

[0084] In this embodiment:

[0085] T1, FLAIR and DIR, referred to as "3 scanning sequences of MRI" or "3 scanning sequences";

[0086] CMIs, short for “cerebral cortical microinfarction”;

[0087] CMIs recognition model, referred to as "AI model or CNN model".

[0088] Attached Figure 1 The diagnostic criteria for cerebral cortical microinfarction proposed in:

[0089] 1. Location: limited to the cerebral cortex;

[0090] 2. Size: 0.5-4mm;

[0091] 3. Signal: T1 low signal, FLAIR high signal, DIR high signal.

[0092] This standard is only a testing standard proposed by the present invention. If there are other standards, they can also be implemented according to the technical principles of the present invention.

[0093] like Figure 3 As shown, according to the clinical diagnostic criteria of CMIs, the cerebral cortex of CMIs patients will show different scanning signal characteristics under the three scanning sequences of MRI.

[0094] Therefore, this solution uses a deep learning model (mainly a convolutional neural network model - CNN model) to learn and train the cerebral cortex scan signal features of CMIs patients under three MRI scan sequences, so as to generate a corresponding AI recognition model. The AI ​​recognition model can recognize the cerebral cortex MRI scan signal features under three scan sequences for clinical CMIs diagnosis.

[0095] First, massive scanning data of patients with cerebral cortical microinfarction under three scanning sequences are collected, and feature engineering is performed. The signal features generated in the feature engineering are learned and trained based on the deep learning model, and the corresponding AI model (CMIs recognition model) is constructed. The AI ​​model is deployed on the background server for MRI image recognition and detection of cerebral cortical microinfarction in clinical patients, so as to quickly identify the signal features on MRI cerebral cortical scan images, assist doctors in quickly making clinical diagnoses of CMIs patients, and adopt a unified and objective detection method to avoid doctors from forming subjective diagnostic results and reduce the clinical requirements for the identification and detection of patients with cerebral cortical microinfarction.

[0096] Therefore, after collecting the cerebral cortical cranial MRI images of clinical patients, the AI ​​model can be used to identify and output the signal features of the graphics under different scanning sequences, so that the background can make a cerebral infarction judgment on the clinical patients. If the cerebral cortical MRI scanning signal features under three different scanning sequences appear on the image, the current clinical patient is deemed to meet the CMIs. Furthermore, the characteristic image recognition and marking under the three scanning sequences are performed on the patient's image, and the CMIs lesions of the clinical patients are further determined on the scanned images through spatial registration and positioning, and the lesions are marked. The lesion location of the CMIs patient can be marked on the cerebral cortical MRI scan graphics, which allows doctors to quickly locate and identify the lesion location of the CMIs patient, thereby improving the detection efficiency of CMIs.

[0097] As a preferred embodiment of the present invention, in step S3, when determining whether the clinical patient meets the CMIs requirement, the following steps are further included:

[0098] If not, further determining the number of CMIs signal features of the clinical diagnostic standard that match the cerebral cortex MRI scan signal features;

[0099] If none of the cerebral cortex MRI scan signal features matches the CMIs signal features of the clinical diagnostic standard, directly outputting the cerebral cortex head MRI image;

[0100] If at least one of the cerebral cortex MRI scan signal features matches the CMIs signal feature of the clinical diagnostic standard, the contour position of the matched cerebral cortex MRI scan signal feature is marked on the cerebral cortex head MRI image of the clinical patient, and the cerebral cortex head MRI image with the marked contour position is output.

[0101] According to the diagnostic criteria for CMIs, CMIs patients need to show three different signal features on MRI images under three scanning sequences. Therefore, when the AI ​​model identifies the corresponding signal features, it is necessary to determine whether the current clinical patient has the signal features required for the three clinical judgment conditions of cerebral cortical microinfarction.

[0102] Therefore, if it is identified that the scan image of the current patient has signal features under three scanning sequences, that is, three different types of signal features, then the current patient is determined to meet the CMIs; if there is only one, two, or none, then the current patient is determined not to meet the judgment criteria of CMIs, and at this time, the current patient is determined not to be CMIs.

[0103] However, in order to allow doctors to conduct further examinations and treatments on patients, this solution will also mark and output the patient's cortical MRI scan signal features identified on the image when it is determined that the patient does not meet the CMIs criteria, so as to facilitate doctors to conduct further diagnosis and treatment of the patient.

[0104] For the specific method of marking the identified signal features on the image (the patient's cerebral cortex head MRI image), please refer to the clinical medicine marking method.

[0105] As a preferred embodiment of the present invention, in step S5, after registering the CMIs lesion areas: C1, C2 and C3 based on the spatial registration method to determine the CMIs lesions of clinical patients, it also includes:

[0106] Acquire a registered cranial MRI image of the CMIs lesion of a clinical patient;

[0107] Based on the three-dimensional reconstruction technology, the cranial MRI image of the CMIs lesion is three-dimensionally reconstructed to construct a three-dimensional CMIs lesion model corresponding to the CMIs lesion;

[0108] Output the three-dimensional CMIs lesion model of clinical patients.

[0109] The head MRI images under different scanning sequences are three-dimensional scanning images themselves. Therefore, they can be combined with three-dimensional scanning for three-dimensional reconstruction and three-dimensional model reconstruction of the marked CMIs lesion area, so as to obtain an intuitive three-dimensional model of the lesion of the CMIs patient, which makes it easier for doctors to intuitively understand the area and size of the patient's CMIs lesion, etc., and to formulate surgical plans better and more accurately.

[0110] Using three-dimensional reconstruction of cranial MRI images can help doctors understand the patient's internal condition more accurately, thereby making more scientific diagnoses and treatment plans.

[0111] The following are the main steps for 3D reconstruction of cranial MRI images:

[0112] 1. Data acquisition: First, two-dimensional image data of the patient's body is obtained through MRI scanning equipment. MRI can generate detailed structural and functional images of the human body and is one of the commonly used imaging technologies in the medical field.

[0113] 2. Image preprocessing: The acquired MRI images may need some preprocessing operations to improve the image quality and the accuracy of subsequent 3D reconstruction. Preprocessing operations may include denoising, contrast enhancement, image registration, etc.

[0114] 3. 3D reconstruction algorithm: Next, the 2D MRI image is converted into a 3D model using a 3D reconstruction algorithm. Commonly used 3D reconstruction algorithms include multi-planar reconstruction (MPR), volume rendering, and surface rendering. These algorithms combine 2D image data into a 3D model in different ways to display the three-dimensional form of the internal structure of the human body.

[0115] Multi-planar reconstruction (MPR): MPR reconstructs the two-dimensional image obtained from the cross-sectional scan into three-dimensional data in voxels, and then reconstructs the coronal, sagittal, and oblique two-dimensional images of any plane based on the volume data. MPR can better display the complex anatomical structure of tissues and organs, which is conducive to the accurate positioning of lesions.

[0116] Volume rendering and surface rendering: Volume rendering and surface rendering are two commonly used direct 3D visualization techniques. Volume rendering generates 3D images by calculating the color and brightness of each voxel, which can show more details and internal structures; while surface rendering generates 3D images by extracting surface information of the area of ​​interest, which has less calculation, but may lose some internal details.

[0117] 4. Post-processing and visualization: After the 3D reconstruction is completed, some post-processing operations may be required, such as smoothing, lighting rendering, etc., to improve the visualization effect of the 3D model. At the same time, the 3D model needs to be displayed to the user in an appropriate way, such as through virtual reality (VR) equipment, 3D printers, etc.

[0118] 5. Application and evaluation: Finally, 3D reconstructed MRI images are used for medical applications, such as auxiliary diagnosis, formulation of surgical plans, etc. At the same time, the accuracy and reliability of 3D reconstruction need to be evaluated to ensure its effectiveness in clinical applications.

[0119] The specific three-dimensional reconstruction is not described in detail in this embodiment, and the user can complete the three-dimensional reconstruction by himself.

[0120] As a preferred embodiment of the present invention, in step S5, after registering the CMIs lesion areas: C1, C2 and C3 based on the spatial registration method to determine the CMIs lesions of clinical patients, it also includes:

[0121] Saving the three-dimensional CMIs lesion model in OBJ format;

[0122] Importing the three-dimensional CMIs lesion model into MeshLab, and using the pyvista library in MeshLab to process and calculate the volume of the three-dimensional CMIs lesion model;

[0123] The volume of the three-dimensional CMIs lesion model is marked on the cerebral cortex cranial MRI image of a clinical patient.

[0124] In order to facilitate doctors to diagnose lesion volume and clinically evaluate the size and volume of lesions (tumors), this program uses a three-dimensional model volume calculation method to calculate the volume of the three-dimensional CMIs lesion model.

[0125] The methods for calculating the volume of a three-dimensional model are:

[0126] Use specialized software or libraries.

[0127] There are many CAD software and libraries (such as Blender, MeshLab, Open3D, etc.) that can calculate the volume of 3D models. These tools usually provide rich functions and options and can handle various complex shapes and models.

[0128] If you are using Python and your model is stored in some format (like STL, OBJ, etc.), consider using a library like pyvista or trimesh to process the model and calculate its volume. Here is an example using pyvista:

[0129] ```python,

[0130] import pyvista as pv,

[0131] Load the model (here we assume there is an STL file named 'model.stl'),

[0132] mesh = pv.read('model.stl');

[0133] Calculating volume

[0134] volume = mesh.volume,

[0135] print(f"The volume of the model is: {volume}").

[0136] like Figure 4 As shown, as a preferred embodiment of the present invention, in step S1, the patient's T1, FLAIR and DIR cranial MRI images are obtained respectively, and the cranial MRI images are used as input to build a corresponding CMIs recognition model based on CNN model training, including:

[0137] The cranial MRI images of several CMIs patients were collected under the MRI scanning sequences: T1, FLAIR and DIR and preprocessed;

[0138] Performing feature engineering on the head MRI images under T1, FLAIR and DIR scanning sequences, respectively extracting the cerebral cortex MRI scanning signal features on the head MRI images under T1, FLAIR and DIR scanning sequences;

[0139] The extracted cerebral cortex MRI scanning signal features are used as input to a preset CNN model, and the cerebral cortex MRI scanning signal features are trained and learned using a convolutional neural network to obtain the CMIs recognition model;

[0140] The CMIs recognition model is verified using the preset cerebral cortex head MRI images under the MRI scanning sequences: T1, FLAIR and DIR;

[0141] After verification, the CMIs recognition model is deployed on the background server.

[0142] Feature engineering is a very important step in image processing, which is used to extract meaningful features from raw images for subsequent analysis or model training. However, it should be noted that traditional feature engineering usually involves manually designing feature extractors, while modern deep learning methods automatically learn image features by training neural networks.

[0143] With the development of deep learning technology, more and more image processing tasks have begun to use deep learning models such as convolutional neural networks (CNNs) to automatically learn image features. This method is usually more effective and flexible than traditional manual feature extraction methods.

[0144] Extracting signal features from MRI (magnetic resonance imaging) scans using convolutional neural networks (CNNs) is a complex but effective process. Here is a detailed explanation of the process:

[0145] 1. MRI image characteristics:

[0146] MRI images provide three-dimensional scans of the body with high definition and no radiation harm to the human body.

[0147] The images contain a lot of structure and information, which are very helpful for diagnosis and treatment decisions.

[0148] However, MRI images also face challenges such as complex noise, high resolution, and large size, which require appropriate preprocessing.

[0149] 2. Convolutional Neural Network (CNN) Features:

[0150] CNN is a type of feedforward neural network that includes convolution calculations and has a deep structure. It is one of the representative algorithms of deep learning.

[0151] It has representation learning capabilities, is capable of translation-invariant classification, and is suitable for pattern recognition tasks such as image and speech.

[0152] The convolution kernel parameter sharing within the hidden layer of CNN and the sparsity of inter-layer connections enable it to learn grid features with less computation.

[0153] 3. MRI image feature extraction based on CNN:

[0154] Preprocessing: Perform denoising, normalization and other preprocessing operations on MRI images to improve image quality and reduce noise interference.

[0155] Convolutional layer: Use multiple convolution kernels to perform convolution operations on the MRI image to extract local features in the image. These convolution kernels can be designed to detect different features such as edges, shapes or textures.

[0156] Pooling layer: Reduce the size of the image by downsampling, reduce the amount of calculation, and retain important feature information. Common pooling operations include maximum pooling and average pooling.

[0157] Fully connected layer: Integrate the features extracted by the convolutional layer and the pooling layer, and perform feature fusion and classification through multiple fully connected layers.

[0158] Optimization and Training: CNN is trained on a large dataset of MRI images using optimization methods such as back-propagation algorithm and gradient descent to optimize network parameters and improve the accuracy of feature extraction.

[0159] Using convolutional neural networks to extract signal features from cranial MRI images is a complex but effective process, which combines the high definition of MRI images with the powerful learning ability of CNN, providing strong support for medical diagnosis and treatment.

[0160] Specifically, the process of training and learning the cerebral cortex MRI scanning signal features using a convolutional neural network to obtain the CMIs recognition model can be implemented in combination with the training and learning principles of CNN.

[0161] Model verification and testing can be performed based on accuracy, sensitivity, AUC curve, etc., which will not be described in detail in this embodiment.

[0162] As a preferred embodiment of the present invention, the S3, comparing the cerebral cortex MRI scan signal characteristics of the clinical patient with the CMIs signal characteristics of the clinical diagnostic standard to determine whether the clinical patient meets the CMIs, comprises:

[0163] Read the cerebral cortex MRI scanning signal characteristics of clinical patients under T1, FLAIR and DIR, and analyze and obtain the position, size and signal characteristics of the cerebral cortex MRI scanning signal characteristics;

[0164] The cerebral cortex MRI scan signal features under T1, FLAIR and DIR are compared with the position, size and signal features of the CMIs signal features of the preset clinical diagnostic standard one by one to determine whether they are consistent:

[0165] If they are consistent, the clinical patient is judged to meet CMIs;

[0166] Otherwise it does not meet the requirements.

[0167] After reading the clinical patient's cerebral cortex scan image, the signal features therein are compared with the signal features in the diagnostic criteria, and the position, size, and signal feature values ​​of the signal features are compared one by one. If they match the signal features in the diagnostic criteria, the patient is identified as a CMIs patient, otherwise not.

[0168] like Figure 5 As shown, as a preferred embodiment of the present invention, S4, if CMIs is met, then according to the contour position of the cerebral cortex MRI scan signal characteristics of the clinical patient, the CMIs lesions at each contour position are located and marked to obtain the corresponding CMIs lesion areas: C1, C2 and C3, including:

[0169] According to the characteristic images shown by the cerebral cortex MRI scanning signal characteristics under T1, FLAIR and DIR, respectively, the image contours displayed by each of the cerebral cortex MRI scanning signal characteristics on the cerebral cortex head MRI image are obtained;

[0170] Based on the same positioning reference, the contours of each image were positioned and marked to obtain the CMIs lesion areas of the cerebral cortex under T1, FLAIR and DIR: C1, C2 and C3.

[0171] Spatial registration methods of lesions are of great significance in the medical field, especially in surgical navigation and lesion localization.

[0172] In the medical field, spatial registration refers to the registration of corresponding points in medical images from different sources (such as CT, MRI, etc.) to obtain the spatial relationship between them.

[0173] For the tracking and positioning of lesions, spatial registration is an important basis for realizing visual medical technology. Through spatial registration, the lesion area determined by the scan can be visualized in the navigation image, helping doctors to more accurately determine the location and boundary of the lesion.

[0174] MRI scanning system, in the display interface of the scanned image, the system has its own positioning reference (a coordinate system can be set).

[0175] The scanning signal characteristics of the patient under different scanning sequences can be used to determine the lesion area and contour under the corresponding scanning sequence according to the position of the characteristic image where each scanning signal characteristic is located. Therefore, the lesion positioning mark can be made on the corresponding scanning image according to the image contour of the characteristic image represented by each signal characteristic. The corresponding lesion area mark can be made on the scanning image under different scanning sequences based on the system default positioning reference.

[0176] Because different scan images are obtained under different scanning sequences, each scan image is positioned according to a unified positioning reference, which facilitates the positioning and marking of the lesion area on each scan image. It is then convenient to fuse and locate the lesion area on each scan image and determine the lesion area of ​​the patient.

[0177] In the coordinate system of the scanning system, the lesion area can be determined by its three-dimensional spatial coordinates based on the characteristic image contour (the three-dimensional coordinates of each pixel point on the edge of the lesion). Therefore, after the system gives the positioning memory, the characteristic image contours on each scanned image will be positioned according to a unified reference, and the corresponding lesion areas will be determined on each scanned image: C1, C2 and C3.

[0178] In order to finally determine the location of the lesion, the lesion regions on the three sequential scanning images are fused to obtain the final lesion location.

[0179] As a preferred embodiment of the present invention, S5, based on the spatial registration method, the CMIs lesion areas: C1, C2 and C3 are registered, the CMIs lesions of clinical patients are determined and marked on the cerebral cortex head MRI image of clinical patients, including:

[0180] The center point of the MRI scanning system is used as a reference;

[0181] Determine respectively the first spatial relationship, the second spatial relationship and the third spatial relationship between the CMIs lesion areas: C1, C2 and C3 and the benchmark;

[0182] The first spatial relationship, the second spatial relationship and the third spatial relationship are integrated to obtain the spatial coordinate relationship of the CMIs lesion of the clinical patient;

[0183] According to the spatial coordinate relationship, the CMIs lesion of the clinical patient is positioned and marked on the cerebral cortex head MRI image of the clinical patient.

[0184] According to the above description, the lesion regions on the scanned images under the three scanning sequences have their own three-dimensional spatial relationships. Therefore, each point on the contour of the lesion region can be fused according to the spatial coordinates to generate the final spatial coordinates of the lesion.

[0185] Specifically, the contour coordinates of the three lesion areas, such as X, Y, and Z, are averaged to generate the X coordinate of the final lesion. The other Y and Z axes are also averaged to generate the three-dimensional spatial coordinates of the final lesion. Through the above average calculation method, the spatial relationship of the lesion areas on the three images can be used to fuse the spatial coordinates of the final patient's cerebral infarction lesion and mark them on the head MRI image.

[0186] By adopting the above-mentioned spatial registration, the lesion areas on the three scanning sequence images are used to register and determine the patient's final cerebral infarction lesion location, which can accurately determine the lesion coordinates and provide more accurate diagnostic results.

[0187] In addition to the above registration, the final coordinate position of the lesion can also be located by the following multiple sets of scan data registration method. The main steps are as follows:

[0188] 1. Data acquisition:

[0189] Use medical imaging equipment (such as CT, MRI, PET-CT, etc.) to scan the patient and obtain multiple sets of original medical imaging data containing lesion information.

[0190] 2. Data preprocessing:

[0191] Perform denoising, enhancement, and standardization on raw medical image data to improve the stability and accuracy of subsequent algorithms.

[0192] 3. Data registration:

[0193] Data registration is a mathematical calculation process that transforms a large set of three-dimensional spatial data points in two or more coordinate systems into a unified coordinate system.

[0194] Control points or targets are set in the scanning area so that there are three or more control points or control targets with the same name on the scanning point cloud of adjacent areas.

[0195] By forcibly attaching control points, adjacent scan data are unified into the same coordinate system.

[0196] The mainstream data registration methods include point set to point set registration method (PSTPS), iterative closest point method, etc.

[0197] During the registration process, it is necessary to solve the rotation matrix R and the three-dimensional translation vector t to describe the transformation relationship between the two coordinate systems.

[0198] 4. Lesion localization technology:

[0199] In the unified coordinate system after registration, the lesion localization technology is used to accurately locate the lesion.

[0200] Common lesion localization technologies include image segmentation technology, feature extraction technology, and three-dimensional reconstruction technology.

[0201] Image segmentation technology divides the images in medical images and separates the lesion area of ​​interest from normal tissue.

[0202] Feature extraction technology extracts key features that can represent lesions from images, such as shape, texture, color, etc.

[0203] Three-dimensional reconstruction technology restores medical imaging data into a three-dimensional model, providing more intuitive and accurate lesion location information.

[0204] 5. Coordinate determination:

[0205] After processing and analysis through the above-mentioned techniques, the final coordinate position of the lesion in the unified coordinate system can be determined.

[0206] This coordinate position can be used for subsequent disease diagnosis, treatment planning, surgical navigation, etc.

[0207] 6. Algorithm optimization:

[0208] Through the above steps, using multiple sets of scanning data registration technology, the final coordinate position of the lesion can be accurately located, providing strong support for clinical diagnosis and treatment.

[0209] Example 2

[0210] Based on the principle of Example 1, on the other hand, a device for automatically detecting cerebral cortical microinfarction is provided, and the device for automatically detecting cerebral cortical microinfarction is used to implement the above-mentioned method for automatically detecting cerebral cortical microinfarction, and the device comprises:

[0211] MRI system, used to collect cerebral cortex cranial MRI images of clinical patients under MRI scanning sequences: T1, FLAIR and DIR, and input them into the backend server;

[0212] Backend server, including:

[0213] HIS system, used to record and save the cerebral cortex head MRI images of clinical patients;

[0214] The CMIs recognition module is used to use a preset CMIs recognition model to recognize and output the cerebral cortex MRI scanning signal characteristics of clinical patients under T1, FLAIR and DIR;

[0215] A feature comparison module, used to compare the cerebral cortex MRI scan signal features of clinical patients with the CMIs signal features of clinical diagnostic standards to determine whether the clinical patients meet the CMIs criteria;

[0216] A lesion marking module is used to locate and mark the CMIs lesions at each contour position according to the contour position of the cerebral cortex MRI scan signal characteristics of the clinical patient if CMIs is met, and obtain the corresponding CMIs lesion areas: C1, C2 and C3;

[0217] A lesion registration module is used to register the CMIs lesion areas: C1, C2 and C3 based on a spatial registration method, determine the CMIs lesions of clinical patients and mark them on the cerebral cortex head MRI image of the clinical patients;

[0218] The output module is used to output the cerebral cortex head MRI image with the CMIs lesions marked.

[0219] The functions and interactions of the above modules are described in conjunction with Example 1 and the attached Figure 3 The corresponding method can be understood and will not be described in detail in this embodiment.

[0220] Example 3

[0221] On the other hand, a detection device is provided, comprising: a processor; a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any one of the automatic detection methods for cerebral cortical microinfarction described in the above-mentioned embodiment 1 is implemented.

[0222] Figure 6 is a schematic diagram of the structure of a detection device provided by an embodiment of the present invention, such as Figure 6 As shown, the detection device may include the above Figure 3 The automatic detection device for cerebral cortical microinfarction is shown.

[0223] Optionally, the detection device 410 may include a first processor 2001 .

[0224] Optionally, the detection device 410 may further include a memory 2002 and a transceiver 2003 .

[0225] The first processor 2001, the memory 2002 and the transceiver 2003 may be connected via a communication bus.

[0226] Combine the following Figure 6 The components of the detection device 410 are described in detail:

[0227] The first processor 2001 is the control center of the detection device 410, and can be a processor or a general term for multiple processing elements. For example, the first processor 2001 is one or more central processing units (CPUs), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (field programmable gate arrays, FPGAs).

[0228] Optionally, the first processor 2001 may perform various functions of the detection device 410 by running or executing a software program stored in the memory 2002 , and calling data stored in the memory 2002 .

[0229] In a specific implementation, as an embodiment, the first processor 2001 may include one or more CPUs, such as Figure 6 CPU0 and CPU1 are shown in FIG.

[0230] In a specific implementation, as an embodiment, the detection device 410 may also include multiple processors, such as Figure 6 The first processor 2001 and the second processor 2004 are shown in FIG. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0231] The memory 2002 is used to store the software program for executing the solution of the present invention, and is controlled to be executed by the first processor 2001. The specific implementation method can refer to the above method embodiment, which will not be repeated here.

[0232] Optionally, the memory 2002 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 2002 may be integrated with the first processor 2001, or may exist independently, and may be accessed through the interface circuit ( Figure 6 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0233] The transceiver 2003 is used to communicate with a network device or a terminal device.

[0234] Optionally, the transceiver 2003 may include a receiver and a transmitter ( Figure 6 The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0235] Optionally, the transceiver 2003 may be integrated with the first processor 2001, or may exist independently and detect the first processor 2001 through the interface circuit ( Figure 6 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0236] It should be noted that Figure 6 The structure of the detection device 410 shown in the figure does not constitute a limitation on the router, and the actual knowledge structure recognition device may include more or fewer components than those shown in the figure, or combine certain components, or arrange the components differently.

[0237] In addition, the technical effects of the detection device 410 can refer to the technical effects of the automatic detection method for cerebral cortical microinfarction described in the above method embodiment, which will not be repeated here.

[0238] It should be understood that the first processor 2001 in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0239] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0240] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0241] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0242] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0243] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0244] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0245] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0246] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0247] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0248] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0249] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0250] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An automatic detection method for cerebral cortical microinfarction, characterized in that: The method comprises: S1, respectively obtain the patient's T1, FLAIR and DIR cranial MRI images, and use the cranial MRI images as input to build a corresponding CMIs recognition model based on CNN model training; S2, collecting cerebral cortex cranial MRI images of clinical patients under MRI scanning sequences: T1, FLAIR and DIR, and inputting them into the CMIs recognition model, and using the CMIs recognition model to recognize and output cerebral cortex MRI scanning signal features under T1, FLAIR and DIR respectively; S3. Compare the cerebral cortex MRI scan signal characteristics of clinical patients with the CMIs signal characteristics of clinical diagnostic standards to determine whether the clinical patients meet the CMIs criteria. The clinical diagnostic criteria are as follows: (1) Location: limited to the cerebral cortex; (2) Size: 0.5-4 mm; (3) Signal: T1 low signal, FLAIR high signal, DIR high signal; S4. If the cerebral cortex MRI scanning signal features under three different scanning sequences appear on the cerebral cortex head MRI image of the clinical patient, it meets the CMIs. According to the contour position of the cerebral cortex MRI scanning signal features of the clinical patient, the CMIs lesions at each contour position are located and marked to obtain the corresponding CMIs lesion areas: C1, C2 and C3; S5. Based on the spatial registration method, the CMIs lesion areas: C1, C2 and C3 are registered to determine the CMIs lesions of the clinical patient and mark them on the cerebral cortex head MRI image of the clinical patient; S6. Output the cerebral cortex cranial MRI image with the CMIs lesions marked.

2. The automatic detection method for cerebral cortical microinfarction according to claim 1, characterized in that: In step S3, when determining whether the clinical patient meets the CMIs, it also includes: If not, further determining the number of CMIs signal features of the clinical diagnostic standard matched by the cerebral cortex MRI scan signal features; If none of the cerebral cortex MRI scan signal features matches the CMIs signal features of the clinical diagnostic standard, directly outputting the cerebral cortex head MRI image; If at least one of the cerebral cortex MRI scan signal features matches the CMIs signal feature of the clinical diagnostic standard, the contour position of the matched cerebral cortex MRI scan signal feature is marked on the cerebral cortex head MRI image of the clinical patient, and the cerebral cortex head MRI image with the marked contour position is output.

3. The automatic detection method for cerebral cortical microinfarction according to claim 1, characterized in that: In step S5, after registering the CMIs lesion areas: C1, C2 and C3 based on the spatial registration method to determine the CMIs lesions of clinical patients, the following steps are further included: Acquire a registered cranial MRI image of the CMIs lesion of a clinical patient; Based on the three-dimensional reconstruction technology, the cranial MRI image of the CMIs lesion is three-dimensionally reconstructed to construct a three-dimensional CMIs lesion model corresponding to the CMIs lesion; Output the three-dimensional CMIs lesion model of clinical patients.

4. The automatic detection method for cerebral cortical microinfarction according to claim 3, characterized in that: In step S5, after registering the CMIs lesion areas: C1, C2 and C3 based on the spatial registration method to determine the CMIs lesions of clinical patients, the following steps are further included: Saving the three-dimensional CMIs lesion model in OBJ format; Importing the three-dimensional CMIs lesion model into MeshLab, and using the pyvista library in MeshLab to process and calculate the volume of the three-dimensional CMIs lesion model; The volume of the three-dimensional CMIs lesion model is marked on the cerebral cortex cranial MRI image of a clinical patient.

5. The automatic detection method for cerebral cortical microinfarction according to claim 1, characterized in that: In step S1, the patient's T1, FLAIR and DIR cranial MRI images are obtained respectively, and the cranial MRI images are used as input to build a corresponding CMIs recognition model based on CNN model training, including: The cranial MRI images of several CMIs patients were collected under the MRI scanning sequences: T1, FLAIR and DIR and preprocessed; Performing feature engineering on the head MRI images under T1, FLAIR and DIR scanning sequences, respectively extracting the cerebral cortex MRI scanning signal features on the head MRI images under T1, FLAIR and DIR scanning sequences; The extracted cerebral cortex MRI scanning signal features are used as input to a preset CNN model, and the cerebral cortex MRI scanning signal features are trained and learned using a convolutional neural network to obtain the CMIs recognition model; The CMIs recognition model is verified using the preset cerebral cortex head MRI images under the MRI scanning sequences: T1, FLAIR and DIR; After verification, the CMIs recognition model is deployed on the background server.

6. The automatic detection method for cerebral cortical microinfarction according to claim 1, characterized in that: In S3, comparing the cerebral cortex MRI scan signal characteristics of the clinical patient with the CMIs signal characteristics of the clinical diagnostic standard to determine whether the clinical patient meets the CMIs requirement includes: Read the cerebral cortex MRI scanning signal characteristics of clinical patients under T1, FLAIR and DIR, and analyze and obtain the position, size and signal characteristics of the cerebral cortex MRI scanning signal characteristics; The cerebral cortex MRI scan signal features under T1, FLAIR and DIR are compared with the position, size and signal features of the CMIs signal features of the preset clinical diagnostic standard one by one to determine whether they are consistent: If they are consistent, the clinical patient is judged to meet CMIs; Otherwise it does not meet the requirements.

7. The automatic detection method for cerebral cortical microinfarction according to claim 4, characterized in that: In S4, if the condition meets the CMIs, the CMIs lesions at each contour position are located and marked according to the contour position of the cerebral cortex MRI scan signal characteristics of the clinical patient, and the corresponding CMIs lesion areas: C1, C2 and C3 are obtained, including: According to the characteristic images shown by the cerebral cortex MRI scanning signal characteristics under T1, FLAIR and DIR, respectively, the image contours displayed by each of the cerebral cortex MRI scanning signal characteristics on the cerebral cortex head MRI image are obtained; Based on the same positioning reference, the contours of each image were positioned and marked to obtain the CMIs lesion areas of the cerebral cortex under T1, FLAIR and DIR: C1, C2 and C3.

8. The automatic detection method for cerebral cortical microinfarction according to claim 7, characterized in that: In S5, based on the spatial registration method, the CMIs lesion areas: C1, C2 and C3 are registered, the CMIs lesions of the clinical patient are determined and marked on the cerebral cortex head MRI image of the clinical patient, including: The center point of the MRI scanning system is used as a reference; Determine respectively the first spatial relationship, the second spatial relationship and the third spatial relationship between the CMIs lesion areas: C1, C2 and C3 and the benchmark; The first spatial relationship, the second spatial relationship and the third spatial relationship are integrated to obtain the spatial coordinate relationship of the CMIs lesion of the clinical patient; According to the spatial coordinate relationship, the CMIs lesion of the clinical patient is positioned and marked on the cerebral cortex head MRI image of the clinical patient.

9. An automatic detection device for cerebral cortical microinfarction, the automatic detection device for cerebral cortical microinfarction is used to implement the automatic detection method for cerebral cortical microinfarction according to any one of claims 1 to 8, characterized in that: The device comprises: MRI system, used to collect cerebral cortex cranial MRI images of clinical patients under MRI scanning sequences: T1, FLAIR and DIR, and input them into the backend server; Backend server, including: HIS system, used to record and save the cerebral cortex head MRI images of clinical patients; The CMIs recognition module is used to use a preset CMIs recognition model to recognize and output the cerebral cortex MRI scanning signal characteristics of clinical patients under T1, FLAIR and DIR; A feature comparison module is used to compare the cerebral cortex MRI scan signal features of clinical patients with the CMIs signal features of clinical diagnostic standards to determine whether the clinical patients meet the CMIs criteria; A lesion marking module is used to locate and mark the CMIs lesions at each contour position according to the contour position of the cerebral cortex MRI scan signal characteristics of the clinical patient if CMIs is met, and obtain the corresponding CMIs lesion areas: C1, C2 and C3; A lesion registration module is used to register the CMIs lesion areas: C1, C2 and C3 based on a spatial registration method, determine the CMIs lesions of clinical patients and mark them on the cerebral cortex head MRI image of the clinical patients; The output module is used to output the cerebral cortex head MRI image with the CMIs lesions marked.

10. A detection device, characterized in that: The detection equipment comprises: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 8 is implemented.

Images