Deep learning-based acute cerebral apoplexy early-stage intelligent screening and early-warning method and system

Abstract

The invention provides an acute cerebral apoplexy early-stage intelligent screening and early warning method and system based on deep learning, and relates to the technical field of medical imaging and artificial intelligence, and the method comprises the steps: obtaining a multi-temporal angiography image of the brain of a patient through a pulse-type dynamic imaging scheme, extracting hemodynamic features through a double-flow feature extraction network, and obtaining a multi-temporal angiography image of the brain of the patient; and constructing a vasculopathy risk distribution diagram by combining a vascular structure analysis model, and marking a potential embolism risk area. Furthermore, in combination with tissue perfusion parameters obtained by multispectral imaging and brain tissue function parameters obtained by electroencephalogram signals, an ischemic penumbra region is identified by using a self-adaptive threshold segmentation model. And finally, based on the ischemic penumbra area, the tissue perfusion parameters and the brain tissue function parameters, calculating a risk level by using a hierarchical decision model with a space-time attention mechanism, and generating a corresponding clinical intervention scheme. According to the invention, the early accurate screening and risk early warning of acute cerebral apoplexy can be realized, and the clinical intervention efficiency is improved.

Description

Technical Field

[0001] The present invention relates to technology, and in particular to an early intelligent screening and warning method and system for acute cerebral stroke based on deep learning. Background Art

[0002] Acute stroke is a neurological emergency that occurs suddenly and has atypical early symptoms. Clinically, it is often diagnosed by imaging methods. Existing imaging analysis methods mainly rely on the subjective judgment of doctors, lack systematic and automated screening and early warning capabilities, and are easily affected by doctors' experience and observation angles, which may lead to inaccurate identification or omission of lesion areas. In addition, traditional imaging analysis technology is difficult to accurately depict the characteristics of cerebral hemodynamics, cannot comprehensively assess the risk level of vascular lesions, and has limited early warning capabilities for potential high-risk patients.

[0003] In recent years, deep learning technology has made significant progress in the field of medical image analysis. Its advantages in feature extraction, pattern recognition and classification have provided new possibilities for early screening of stroke. However, existing studies have mostly focused on static image analysis, lacking in-depth exploration of hemodynamic characteristics, and unable to effectively combine vascular structure, blood flow timing and functional parameters for comprehensive evaluation. In addition, the current intelligent analysis system has not yet formed a complete chain from image data acquisition, vascular lesion risk analysis to intervention plan recommendation, making it difficult to meet the clinical needs of rapid screening and early intervention of acute stroke.

[0004] Therefore, there is an urgent need to propose an intelligent analysis method that integrates vascular structure, blood flow characteristics and functional parameters to achieve automated, high-precision stroke risk assessment and early warning. Summary of the invention

[0005] The embodiments of the present invention provide a method and system for early intelligent screening and warning of acute stroke based on deep learning, which can solve the problems in the prior art.

[0006] According to a first aspect of the embodiments of the present invention, Provide an early intelligent screening and warning method for acute stroke based on deep learning, including: A pulsed dynamic imaging scheme is used to trigger a scanning sequence within the cardiac cycle to acquire multi-phase angiographic images, and a hemodynamic analysis is performed on the multi-phase angiographic images to obtain blood flow velocity vector field data. A blood flow feature tensor is constructed based on the blood flow velocity vector field data, and the blood flow feature tensor is input into a dual-flow feature extraction network, wherein the vascular structure features are extracted through a spatial feature branch, and the blood flow time series features are extracted through a time series feature branch, and the vascular structure features and the blood flow time series features are fused to generate a hemodynamic feature map; The hemodynamic characteristic map is processed by a directional filter to obtain an enhanced vascular image, and the vascular topology data is reconstructed using a minimum spanning tree algorithm. The vascular topology data is input into a recursive neural tensor network to calculate the risk value of vascular branch nodes to generate a vascular lesion risk distribution map, and potential embolism risk areas are marked based on the vascular lesion risk distribution map. A multispectral imaging device is used to obtain tissue perfusion parameters in potential embolism risk areas, and electroencephalogram signals in potential embolism risk areas are collected to obtain brain tissue function parameters. The tissue perfusion parameters and brain tissue function parameters are adaptively thresholded to obtain the ischemic penumbra area, and the area value of the ischemic penumbra area is calculated. The area value of the ischemic penumbra area, tissue perfusion parameters, and brain tissue function parameters are input into a hierarchical decision model to calculate the risk level value, and a clinical intervention plan is generated according to the risk level value.

[0007] In an optional embodiment, The pulsed dynamic imaging scheme is used to trigger the scanning sequence to acquire multi-phase angiographic images during the cardiac cycle, and the multi-phase angiographic images are subjected to hemodynamic analysis to obtain blood flow velocity vector field data. The blood flow characteristic tensor is constructed based on the blood flow velocity vector field data, including: The patient's electrocardiogram (ECG) signal is collected to construct a pulsed dynamic imaging scheme, wherein the pulsed dynamic imaging scheme uses a dual-domain sparse sampling model in the frequency domain space and the time domain for undersampling, and multiple trigger windows are constructed within the cardiac cycle based on the patient's ECG signal. A spiral angle incremental ray sampling strategy is used within the multiple trigger windows to trigger a scanning sequence, thereby obtaining non-uniform grid data for multi-phase angiography; Iteratively reconstructing the non-uniform grid data, optimizing the data consistency in the spatial domain and the temporal domain in each iteration, and finally obtaining a multi-phase angiography image with high temporal and spatial resolution; Inputting the multi-phase angiography image into a multi-scale hemodynamic feature extraction network, the multi-scale hemodynamic feature extraction network includes an adaptive deformation convolution layer, a spatial attention module, and a cyclic feature extraction unit, wherein the adaptive deformation convolution layer extracts local blood flow features to obtain vascular feature data, the spatial attention module determines the vascular branch connection relationship based on the vascular feature data to obtain vascular network data, and the cyclic feature extraction unit performs time series analysis on the vascular network data based on pressure wave propagation constraints, and finally obtains blood flow velocity vector field data; Phase enhancement technology is used to calculate the phase difference information between adjacent phase images that characterizes the magnitude and direction of blood flow velocity. Based on the phase difference information, nonlinear correction is performed on the blood flow velocity vector field data. According to the corrected blood flow velocity vector field data, a blood flow characteristic tensor including blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics is constructed.

[0008] In an optional embodiment, The blood flow feature tensor is input into the dual-flow feature extraction network, wherein the vascular structure features are extracted through the spatial feature branch, and the blood flow timing features are extracted through the timing feature branch. The vascular structure features and the blood flow timing features are fused to generate the hemodynamic feature map, including: Inputting the blood flow feature tensor into a dual-flow feature extraction network, the dual-flow feature extraction network includes a spatial feature branch and a temporal feature branch, the blood flow feature tensor includes blood flow vortex distribution features, blood vessel wall stress distribution features, blood flow complexity features and pressure wave propagation features; The spatial feature branch generates a convolution kernel deformation parameter based on the vascular morphology information in the blood flow feature tensor, performs multi-scale convolution operations on the deformable convolution kernel group and the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics to obtain a local feature atlas, and performs channel dimension weighting on the local feature atlas to obtain vascular structure characteristics; The time series feature branch performs phase analysis on the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics in the blood flow feature tensor, extracts the instantaneous phase and instantaneous amplitude of each characteristic component through Hilbert transform, calculates the phase difference between time adjacent windows to obtain a phase change curve group, constructs a phase consistency matrix based on the phase change curve group, selectively enhances the instantaneous amplitude and reconstructs it to obtain the blood flow time series characteristics; The mutual information between the vascular structure features and the blood flow timing features is calculated to obtain a feature association matrix, a bidirectional attention weight is constructed based on the feature association matrix, the vascular structure features and the blood flow timing features are adaptively enhanced respectively, and the enhanced features are fused through a jump connection structure to obtain a hemodynamic feature map.

[0009] In an optional embodiment, The hemodynamic characteristic map is processed by a directional filter to obtain an enhanced vascular image, and the minimum spanning tree algorithm is used to reconstruct the vascular topology data including: Extracting blood vessel morphological features in the hemodynamic characteristic map, determining the scale range of a Gaussian second-order derivative filter based on the blood vessel morphological features, and rotating the Gaussian second-order derivative filter within a preset angle range to generate a directional filter; Filtering the hemodynamic characteristic map using the directional filter to obtain multi-scale directional response values, determining weighting coefficients according to the direction of blood vessels in the hemodynamic characteristic map, weighting the multi-scale directional response values, obtaining the weighted maximum response value and the corresponding optimal direction parameters and optimal scale parameters, and generating an enhanced blood vessel image; Performing local maximum suppression processing on the enhanced vascular image, obtaining a set of candidate points for the vascular centerline, constructing an undirected graph, processing the undirected graph using a minimum spanning tree algorithm, determining a dynamic weight threshold according to the vascular density in the enhanced vascular image, removing node connections greater than the dynamic weight threshold, and obtaining an initial vascular topology structure; On the path corresponding to the connection edge of the initial vascular topological structure, blood flow vortex distribution, vascular wall stress distribution, blood flow complexity and pressure wave propagation characteristics are extracted, and the gradient change value is calculated to construct a feature consistency index; The connection relationship in the initial vascular topology structure is verified according to the feature consistency index, and the connection relationship with a gradient change value greater than a preset gradient threshold is identified as abnormal and corrected or deleted, so as to complete the vascular connectivity reconstruction and generate vascular topology structure data.

[0010] In an optional embodiment, The vascular topology data is input into the recursive neural tensor network to calculate the risk value of the vascular branch node to generate a vascular lesion risk distribution map. Based on the vascular lesion risk distribution map, the potential embolism risk areas are marked, including: Extracting feature vectors of vascular branch nodes from vascular topological structure data, calculating a topological relationship matrix between nodes, wherein the topological relationship matrix records the superior-subordinate and the same-level connection relationships of nodes, and constructing a tree-shaped computational graph based on the feature vectors and the topological relationship matrix, wherein each node in the graph contains corresponding feature vectors and topological relationship data; Inputting the tree-shaped computation graph into a recursive neural tensor network, wherein the recursive neural tensor network comprises a third-order tensor computation unit, wherein the third-order tensor computation unit performs a tensor product operation on feature vectors of adjacent nodes to obtain structural correlation data between nodes and node feature correlation strength; Based on the inter-node structural correlation data and the node feature association strength, an attention weight matrix is ​​constructed, wherein the attention weight matrix includes a feature dimension weight component and a node association weight component, and the feature vector is weighted according to the feature dimension weight component and the node association weight component to generate a weighted feature vector that integrates local structural information; The weighted feature vector is mapped using a nonlinear activation function to obtain an initial risk assessment value, and the initial risk assessment value is input into a recursive calculation module. The recursive calculation module calculates from the terminal branch node in a bottom-up manner, and fuses the feature information and risk information of the lower-level node based on the feature dimension weight component and the node association weight component, and updates the risk assessment value of the upper-level node until all nodes are traversed to generate a final risk assessment value; The final risk assessment value is normalized to generate a risk distribution map, the vascular structure is partitioned according to the risk value in the risk distribution map, and the area with a risk value higher than a set risk threshold is marked as a potential embolism risk area.

[0011] In an optional embodiment, Multispectral imaging equipment is used to obtain tissue perfusion parameters in potential embolic risk areas, and electroencephalogram signals in potential embolic risk areas are collected to obtain brain tissue function parameters. The tissue perfusion parameters and brain tissue function parameters are segmented by adaptive thresholds to obtain the ischemic penumbra area, including: A multispectral imaging device is used to perform continuous wavelength scanning within a preset wavelength range through a liquid crystal tunable filter, and multiple frames of images are accumulated and collected at each wavelength point, and the average value is taken to obtain a spectral data cube; A tissue optical parameter inversion model is constructed according to the incident light intensity, the transmitted light intensity and the tissue thickness, the spectral data cube is input into the tissue optical parameter inversion model, the oxygenated hemoglobin concentration and the deoxygenated hemoglobin concentration are calculated by the least square fitting method, the tissue blood oxygen saturation is calculated by the ratio of the oxygenated hemoglobin concentration to the sum of the oxygenated hemoglobin concentration and the deoxygenated hemoglobin concentration, and the tissue reflectance ratio is calculated at a preset characteristic wavelength to obtain the tissue perfusion index; The EEG signal of the potential embolism risk area is collected and band-pass filtered, and the filtered EEG signal is decomposed into a plurality of preset frequency bands by using a wavelet packet decomposition method, and the ratio of the energy of each preset frequency band to the total energy is calculated to obtain a relative power spectrum, and the filtered EEG signal is converted into a symbol sequence, and then the probability distribution of different arrangement patterns is statistically obtained to obtain an arrangement entropy value, and the relative power spectrum and the arrangement entropy value are combined in a preset order to obtain brain tissue function parameters; The tissue perfusion parameters and brain tissue function parameters are combined into a feature vector, the Euclidean distance from each sample point in the feature vector to each cluster center is calculated, the membership of each sample point to each cluster center is calculated based on the Euclidean distance, the cluster center is updated according to the membership, and the iteration is repeated until the difference between the objective functions of two adjacent iterations is less than a preset convergence threshold to obtain the final cluster center; The feature vector is classified based on the final cluster center to obtain the spatial distribution of ischemic area, penumbra area and normal area, a set of pixel points in the penumbra area is extracted for edge detection to obtain a closed boundary contour, and the area within the closed boundary contour is marked as the ischemic penumbra area.

[0012] In an optional embodiment, The area value of the ischemic penumbra region is calculated, and the area value of the ischemic penumbra region, tissue perfusion parameters, and brain tissue function parameters are input into the hierarchical decision model to calculate the risk level value. The clinical intervention plan generated according to the risk level value includes: Marking the connected domains of the ischemic penumbra area and constructing a neuron connection map, calculating the topological relationship matrix between neurons, counting the total number of pixels in the ischemic penumbra area, and converting the total number of pixels into an area value of the ischemic penumbra area according to a spatial resolution parameter; A hierarchical decision model of the spatiotemporal attention mechanism is constructed, including a spatial attention module and a temporal attention module; the spatial attention module uses a wavelet multi-resolution analysis method to decompose the tissue perfusion parameters and the topological relationship matrix into spatial feature matrices of multiple scales, and extracts the dynamic change characteristics of the tissue perfusion parameters through a recursive neural network to calculate the spatial weight coefficient; the temporal attention module uses a nonlinear time series decomposition method to decompose the brain tissue function parameters into feature sequences of different time scales, and extracts the time series change characteristics through an attention autoencoder to calculate the time weight coefficient; A multimodal risk assessment feature set is constructed according to the area value of the ischemic penumbra region, the weighted tissue perfusion parameter, and the weighted brain tissue function parameter, and the fused feature vector is input into a graph convolutional neural network together with a neuron connection map to obtain a fused feature vector, and the fused feature vector is input into a residual network to obtain a risk level value; The risk evolution trend is predicted based on the risk level value, and clinical intervention rules are selected in combination with historical intervention effect data to generate a clinical intervention decision tree. The decision tree nodes are prioritized based on the risk evolution trend, and intervention effect and resource consumption are used as evaluation indicators to obtain a clinical intervention plan with priority.

[0013] According to a second aspect of the embodiments of the present invention, Provide an early intelligent screening and warning system for acute stroke based on deep learning, including: The first unit is used to trigger a scanning sequence to acquire multi-phase angiography images within a cardiac cycle through a pulsed dynamic imaging scheme, perform hemodynamic analysis on the multi-phase angiography images to obtain blood flow velocity vector field data, construct a blood flow feature tensor according to the blood flow velocity vector field data, input the blood flow feature tensor into a dual-flow feature extraction network, wherein the vascular structure feature is extracted through a spatial feature branch, the blood flow time series feature is extracted through a time series feature branch, and the vascular structure feature is fused with the blood flow time series feature to generate a hemodynamic feature map; The second unit is used to process the hemodynamic characteristic map through a directional filter to obtain an enhanced vascular image, reconstruct the vascular topology data using a minimum spanning tree algorithm, input the vascular topology data into a recursive neural tensor network to calculate the risk value of the vascular branch node to generate a vascular lesion risk distribution map, and mark the potential embolism risk area based on the vascular lesion risk distribution map; The third unit is used to obtain tissue perfusion parameters of potential embolic risk areas using multispectral imaging equipment, collect electroencephalogram signals of potential embolic risk areas to obtain brain tissue function parameters, perform adaptive threshold segmentation on tissue perfusion parameters and brain tissue function parameters to obtain the ischemic penumbra area, calculate the area value of the ischemic penumbra area, input the area value of the ischemic penumbra area, tissue perfusion parameters and brain tissue function parameters into a hierarchical decision model to calculate the risk level value, and generate a clinical intervention plan based on the risk level value.

[0014] According to a third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0015] According to a fourth aspect of the embodiments of the present invention, A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0016] In this embodiment, the pulsed dynamic imaging scheme and hemodynamic analysis can be used to more accurately capture the changes in cerebral blood flow, and analyze the vascular structure and timing characteristics, so as to more accurately identify potential embolic risk areas and achieve early warning of stroke. Compared with traditional methods, the present invention can detect signs of stroke earlier and improve the efficiency of early diagnosis. Through the vascular structure analysis model and the recursive neural tensor network, the risk of vascular lesions can be quantitatively evaluated and a vascular lesion risk distribution map can be generated. Combined with the tissue perfusion parameters and brain tissue function parameters obtained by multispectral imaging and electroencephalography technology, the present invention can more comprehensively evaluate the damage of the ischemic penumbra area, so as to more accurately evaluate the risk level of stroke. According to the area value, tissue perfusion parameters, brain tissue function parameters and risk level values ​​of the ischemic penumbra area, a personalized clinical intervention plan can be generated, and the priority of the plan can be adjusted according to the risk level. This personalized intervention strategy can better guide clinical treatment, improve treatment effects, and gain precious rescue time for patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1This is a flow chart of an early intelligent screening and warning method for acute stroke based on deep learning according to an embodiment of the present invention; Figure 2 This is a comparison chart of the accuracy of blood vessel connectivity reconstruction in an embodiment of the present invention; Figure 3 This is a diagram showing the effect of vascular risk distribution according to an embodiment of the present invention; Figure 4 A schematic diagram of the structure of an early intelligent screening and warning system for acute stroke based on deep learning in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0020] Figure 1 FIG. 1 is a flow chart of an early intelligent screening and warning method for acute stroke based on deep learning according to an embodiment of the present invention. Figure 1 As shown, the method includes: A pulsed dynamic imaging scheme is used to trigger a scanning sequence within the cardiac cycle to acquire multi-phase angiographic images, and a hemodynamic analysis is performed on the multi-phase angiographic images to obtain blood flow velocity vector field data. A blood flow feature tensor is constructed based on the blood flow velocity vector field data, and the blood flow feature tensor is input into a dual-flow feature extraction network, wherein the vascular structure features are extracted through a spatial feature branch, and the blood flow time series features are extracted through a time series feature branch, and the vascular structure features and the blood flow time series features are fused to generate a hemodynamic feature map; The hemodynamic characteristic map is processed by a directional filter to obtain an enhanced vascular image, and the vascular topology data is reconstructed using a minimum spanning tree algorithm. The vascular topology data is input into a recursive neural tensor network to calculate the risk value of vascular branch nodes to generate a vascular lesion risk distribution map, and potential embolism risk areas are marked based on the vascular lesion risk distribution map. A multispectral imaging device is used to obtain tissue perfusion parameters in potential embolism risk areas, and electroencephalogram signals in potential embolism risk areas are collected to obtain brain tissue function parameters. The tissue perfusion parameters and brain tissue function parameters are adaptively thresholded to obtain the ischemic penumbra area, and the area value of the ischemic penumbra area is calculated. The area value of the ischemic penumbra area, tissue perfusion parameters, and brain tissue function parameters are input into a hierarchical decision model to calculate the risk level value, and a clinical intervention plan is generated according to the risk level value.

[0021] In an optional embodiment, a pulsed dynamic imaging scheme is used to trigger a scanning sequence within a cardiac cycle to acquire multi-phase angiographic images, a hemodynamic analysis is performed on the multi-phase angiographic images to obtain blood flow velocity vector field data, and a blood flow characteristic tensor is constructed according to the blood flow velocity vector field data, including: The patient's electrocardiogram (ECG) signal is collected to construct a pulsed dynamic imaging scheme, wherein the pulsed dynamic imaging scheme uses a dual-domain sparse sampling model in the frequency domain space and the time domain for undersampling, and multiple trigger windows are constructed within the cardiac cycle based on the patient's ECG signal. A spiral angle incremental ray sampling strategy is used within the multiple trigger windows to trigger a scanning sequence, thereby obtaining non-uniform grid data for multi-phase angiography; Iteratively reconstructing the non-uniform grid data, optimizing the data consistency in the spatial domain and the temporal domain in each iteration, and finally obtaining a multi-phase angiography image with high temporal and spatial resolution; Inputting the multi-phase angiography image into a multi-scale hemodynamic feature extraction network, the multi-scale hemodynamic feature extraction network includes an adaptive deformation convolution layer, a spatial attention module, and a cyclic feature extraction unit, wherein the adaptive deformation convolution layer extracts local blood flow features to obtain vascular feature data, the spatial attention module determines the vascular branch connection relationship based on the vascular feature data to obtain vascular network data, and the cyclic feature extraction unit performs time series analysis on the vascular network data based on pressure wave propagation constraints, and finally obtains blood flow velocity vector field data; Phase enhancement technology is used to calculate the phase difference information between adjacent phase images that characterizes the magnitude and direction of blood flow velocity. Based on the phase difference information, nonlinear correction is performed on the blood flow velocity vector field data. According to the corrected blood flow velocity vector field data, a blood flow characteristic tensor including blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics is constructed.

[0022] Exemplarily, the patient's ECG signal is first collected, and the signal includes a front wave signal, a ventricular depolarization wave group signal and an after wave signal. A pulsed dynamic imaging scheme is constructed using the ECG signal. The scheme uses a dual-domain sparse sampling model in the frequency domain space and time domain for undersampling to reduce the amount of data collection. According to the front wave, ventricular depolarization wave group and after wave in the ECG signal, multiple trigger windows are set in the cardiac cycle. In each trigger window, a spiral angle incremental ray sampling strategy is used to trigger the scanning sequence to collect data. For example, in one cardiac cycle, the front wave is the first trigger window, the ventricular depolarization wave group is the second trigger window, and the after wave is the third trigger window. In the first trigger window, the spiral ray angle is 0 degrees to start collecting; in the second trigger window, the spiral ray angle is 120 degrees to start collecting; in the third trigger window, the spiral ray angle is 240 degrees to start collecting. This sampling strategy can obtain non-uniform grid data of multi-phase angiography, for example, non-uniform grid data of three phases are collected in one cardiac cycle, and the data of each phase contains projection data of different spiral angles.

[0023] The acquired non-uniform grid data is input into the directional transformation iterative reconstruction model for reconstruction. The model contains four main modules: a non-uniform Fourier operation module for processing non-uniform sampling data, a directional transformation module for extracting vascular edge and direction information, a time gradient calculation module for constructing time continuity constraints, and a matrix decomposition module for reducing image noise. Directional transformations, for example, use curve transformations (Curvelet transformation) or shear transformations (Shearlet transformation). The time gradient calculation module constructs time continuity constraints by calculating the pixel value differences between adjacent phase images. The matrix decomposition module, for example, uses principal component analysis or singular value decomposition. The non-uniform grid data is iteratively reconstructed through this model. In each iteration, the data consistency of the spatial domain and the time domain is optimized simultaneously, and finally a multi-phase angiography image with high spatiotemporal resolution is obtained. For example, after iterative reconstruction, three phase angiography images are obtained, each with a spatial resolution of 512x512 pixels and a temporal resolution of three different time points in the cardiac cycle.

[0024] The reconstructed multi-phase angiography images are input into the multi-scale hemodynamic feature extraction network. The network includes an adaptive deformation convolution layer, a spatial attention module, and a cyclic feature extraction unit. The adaptive deformation convolution layer extracts local blood flow features to obtain vascular feature data. The spatial attention module determines the vascular branch connection relationship based on the vascular feature data to obtain vascular network data. The cyclic feature extraction unit performs time series analysis on the vascular network data based on the pressure wave propagation constraint, and finally obtains the blood flow velocity vector field data. For example, the adaptive deformation convolution layer extracts the diameter, curvature and other features of the blood vessel, the spatial attention module identifies the vascular branch points and connection relationships, and the cyclic feature extraction unit analyzes the changes in the vascular network data in different phases, and finally obtains the blood flow velocity size and direction of each pixel.

[0025] Phase enhancement technology is used to calculate the phase difference information between adjacent phase images that characterizes the magnitude and direction of blood flow velocity. Based on the phase difference information, nonlinear correction is performed on the blood flow velocity vector field data to eliminate phase aliasing. A blood flow feature tensor containing blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics, and pressure wave propagation characteristics is constructed based on the corrected blood flow velocity vector field data. For example, the blood flow vortex distribution characteristics are calculated based on the curl of the blood flow velocity vector field, the vascular wall stress distribution characteristics are calculated based on the blood flow velocity gradient, the blood flow complexity characteristics are calculated based on the degree of change of the blood flow velocity vector field, and the pressure wave propagation characteristics are calculated based on the changes in the blood flow velocity vector field in different phases.

[0026] In this embodiment, the pulsed dynamic imaging scheme can accurately trigger the scanning sequence within the cardiac cycle to obtain multi-phase angiography images with high temporal and spatial resolution; the blood flow velocity vector field data is extracted through hemodynamic analysis, thereby achieving in-depth analysis of blood flow characteristics. The dual-domain sparse sampling and iterative reconstruction model are used to optimize data consistency and improve image quality; the multi-scale hemodynamic feature extraction network is used to effectively extract vascular characteristics and blood flow information, and further accurately construct the blood flow feature tensor. The application of phase enhancement technology eliminates phase aliasing, improves the accuracy of the blood flow velocity vector field, and provides a more accurate analysis tool for the diagnosis and treatment of vascular diseases.

[0027] In an optional embodiment, the blood flow feature tensor is input into a dual-flow feature extraction network, wherein the vascular structure feature is extracted through the spatial feature branch, the blood flow timing feature is extracted through the timing feature branch, and the vascular structure feature and the blood flow timing feature are fused to generate a hemodynamic feature map, including: Inputting the blood flow feature tensor into a dual-flow feature extraction network, the dual-flow feature extraction network includes a spatial feature branch and a temporal feature branch, the blood flow feature tensor includes blood flow vortex distribution features, blood vessel wall stress distribution features, blood flow complexity features and pressure wave propagation features; The spatial feature branch generates a convolution kernel deformation parameter based on the vascular morphology information in the blood flow feature tensor, performs multi-scale convolution operations on the deformable convolution kernel group and the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics to obtain a local feature atlas, and performs channel dimension weighting on the local feature atlas to obtain vascular structure characteristics; The time series feature branch performs phase analysis on the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics in the blood flow feature tensor, extracts the instantaneous phase and instantaneous amplitude of each characteristic component through Hilbert transform, calculates the phase difference between time adjacent windows to obtain a phase change curve group, constructs a phase consistency matrix based on the phase change curve group, selectively enhances the instantaneous amplitude and reconstructs it to obtain the blood flow time series characteristics; The mutual information between the vascular structure features and the blood flow timing features is calculated to obtain a feature association matrix, a bidirectional attention weight is constructed based on the feature association matrix, the vascular structure features and the blood flow timing features are adaptively enhanced respectively, and the enhanced features are fused through a jump connection structure to obtain a hemodynamic feature map.

[0028] Exemplarily, first, a blood flow feature tensor is obtained. This tensor contains four key blood flow information dimensions: blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics, and pressure wave propagation characteristics. For example, computational fluid dynamics simulation technology can be used to simulate a specific blood vessel model to obtain velocity field, pressure field and other data at different time points and spatial positions. Then, based on these data, characteristics such as blood flow vortex, vascular wall stress, blood flow complexity and pressure wave propagation are calculated, and these characteristics are organized into a multidimensional tensor. Assuming that the time sampling points are 100 and the space sampling points are 256x256, the blood flow feature tensor dimension can be 100x256x256x4.

[0029] The blood flow feature tensor is input into a dual-stream feature extraction network. The network consists of a spatial feature branch and a temporal feature branch, which are used to extract vascular structure features and blood flow temporal features, respectively.

[0030] In the spatial feature branch, the deformation parameters of the deformable convolution kernel are first generated using the vascular morphology information in the blood flow feature tensor. Specifically, the sampling point position offset is calculated based on the vascular cross-sectional change rate and the local curvature. For example, the sampling point position offset is also larger in the area with a larger vascular cross-sectional change rate, so that the convolution kernel can pay more attention to the changes in vascular morphology. Assuming that in a local area, the diameter of the vascular cross-sectional area changes from 2 pixels to 4 pixels, the sampling point position offset of the area can be set to 1 pixel. The calculation method of the local curvature is similar. The larger the curvature, the larger the sampling point position offset. Then, the standard convolution kernel is adaptively modulated according to these deformation parameters to obtain a deformable convolution kernel group. For example, after a 3x3 standard convolution kernel is modulated by the deformation parameters, the position of its sampling points may no longer be a regular grid distribution, but adaptively adjusted according to the vascular morphology. The obtained deformable convolution kernel group is subjected to multi-scale convolution operations with the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics, and pressure wave propagation characteristics to obtain a local feature atlas. Finally, the local feature atlas is weighted by channel dimension to obtain the vascular structure features. For example, different weights can be assigned according to the importance of different features, such as the weight of blood flow vortex distribution feature is 0.4, the weight of vascular wall stress distribution feature is 0.3, the weight of blood flow complexity feature is 0.2, and the weight of pressure wave propagation feature is 0.1.

[0031] In the time series feature branch, phase analysis is performed on the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics in the blood flow feature tensor. The instantaneous phase and instantaneous amplitude of each characteristic component are extracted by Hilbert transform. Then, the phase difference between adjacent windows in time is calculated to obtain a phase change curve group. For example, the phase difference between each two adjacent time points can be calculated to obtain a series of phase change curves. A phase consistency matrix is ​​constructed based on the phase change curve group, which is used to characterize the time periodicity of each characteristic component. For example, if the phase change curves of two characteristic components are highly similar, their corresponding element values ​​in the phase consistency matrix are larger. The instantaneous amplitude is selectively enhanced and reconstructed using the phase consistency matrix to obtain the blood flow time series characteristics. For example, for characteristic components with higher phase consistency, their instantaneous amplitude will be enhanced, while for characteristic components with lower phase consistency, their instantaneous amplitude will be weakened.

[0032] The mutual information between the vascular structure features and the blood flow timing features is calculated to obtain the feature association matrix. Based on the feature association matrix, a bidirectional attention weight is constructed, which is used to characterize the degree of mutual enhancement between the structural features and the timing features. For example, if there is a strong correlation between the vascular structure features and the blood flow timing features, the corresponding attention weight is larger.

[0033] The bidirectional attention weights are used to adaptively enhance the vascular structure features and blood flow timing features. For example, the attention weights are multiplied by the corresponding features to achieve adaptive enhancement of the features. The enhanced features are fused through the jump connection structure to obtain the final hemodynamic feature map.

[0034] In this embodiment, the existing hemodynamic feature extraction method usually relies on a fixed convolution kernel for feature capture, which is difficult to adapt to the influence of complex vascular morphology on blood flow characteristics, resulting in limited accuracy of spatial feature extraction, and limiting the integrity and accuracy of the final feature expression. The present application optimizes the spatial information and temporal information of the blood flow feature tensor through a dual-stream feature extraction network. In the process of spatial feature extraction, the convolution kernel deformation parameters are generated based on the vascular morphology information, so that the deformable convolution kernel adapts to the local structure of different blood flow characteristics, achieves more accurate feature capture, and enhances key vascular structure features through channel dimension weighting. In terms of temporal feature extraction, the instantaneous phase and instantaneous amplitude of the blood flow characteristics are analyzed by Hilbert transform, and a phase consistency matrix is ​​constructed to achieve selective enhancement of blood flow timing features and improve the expression ability of timing information. The mutual information between vascular structure features and blood flow timing features is calculated through the feature association matrix, and adaptive enhancement is performed using a bidirectional attention mechanism, and features are fused through a jump connection structure to achieve a more discriminative hemodynamic feature representation. Compared with the existing technology, this application scheme introduces deformable convolution to adapt to the local changes of different blood flow characteristics, improves the accuracy of spatial features, and uses instantaneous phase analysis methods to enhance the ability to express temporal features. In addition, the innovative introduction of feature association matrix and bidirectional attention mechanism enables a closer integration of spatial and temporal features, improving the overall feature expression effect. It can extract hemodynamic features more comprehensively and accurately, providing more reliable data support for hemodynamic analysis.

[0035] In an optional embodiment, the hemodynamic characteristic map is processed by a directional filter to obtain an enhanced vascular image, and the vascular topology data is reconstructed by using a minimum spanning tree algorithm, including: Extracting blood vessel morphological features in the hemodynamic characteristic map, determining the scale range of a Gaussian second-order derivative filter based on the blood vessel morphological features, and rotating the Gaussian second-order derivative filter within a preset angle range to generate a directional filter; Filtering the hemodynamic characteristic map using the directional filter to obtain multi-scale directional response values, determining weighting coefficients according to the direction of blood vessels in the hemodynamic characteristic map, weighting the multi-scale directional response values, obtaining the weighted maximum response value and the corresponding optimal direction parameters and optimal scale parameters, and generating an enhanced blood vessel image; Performing local maximum suppression processing on the enhanced vascular image, obtaining a set of candidate points for the vascular centerline, constructing an undirected graph, processing the undirected graph using a minimum spanning tree algorithm, determining a dynamic weight threshold according to the vascular density in the enhanced vascular image, removing node connections greater than the dynamic weight threshold, and obtaining an initial vascular topology structure; On the path corresponding to the connection edge of the initial vascular topological structure, blood flow vortex distribution, vascular wall stress distribution, blood flow complexity and pressure wave propagation characteristics are extracted, and the gradient change value is calculated to construct a feature consistency index; The connection relationship in the initial vascular topology structure is verified according to the feature consistency index, and the connection relationship with a gradient change value greater than a preset gradient threshold is identified as abnormal and corrected or deleted, so as to complete the vascular connectivity reconstruction and generate vascular topology structure data.

[0036] Exemplarily, first, a hemodynamic characteristic map is obtained. The hemodynamic characteristic map can be obtained, for example, by computational fluid dynamics (CFD) simulation or image-based methods (such as phase contrast magnetic resonance imaging). Taking CFD simulation as an example, by establishing a patient-specific vascular geometry model, setting the physical properties of the blood (such as viscosity, density), and defining boundary conditions (such as inlet velocity, outlet pressure), simulation calculations are performed to obtain a hemodynamic characteristic map containing information such as velocity, pressure, and vorticity. Assume that the resolution of the obtained hemodynamic characteristic map is 512x512 pixels, and the value range is 0 to 1, indicating the magnitude of the blood flow velocity.

[0037] Next, the hemodynamic feature map is input into the vascular structure analysis model. The core of this model is to enhance the vascular image using a directional filter and reconstruct the vascular connectivity using the minimum spanning tree algorithm. The vascular structure analysis model first analyzes the vascular morphological features in the hemodynamic feature map, such as the vascular radius. Assume that the vascular radius obtained by the image processing algorithm varies from 2 to 10 pixels. Based on this range, the scale range of the Gaussian second-order derivative filter is determined to be 2 to 10 pixels to match the size of the blood vessel.

[0038] Then, a directional filter is generated. The Gaussian second-order derivative filter is rotated at a fixed interval (e.g., 1 degree) within a preset angle range (e.g., 0 to 180 degrees) to generate multiple directional filters. Each filter corresponds to a specific direction and scale. For example, 180 directions are generated, each direction corresponds to a scale of 2 to 10 pixels, and a total of 180*9=1620 filters are generated.

[0039] Afterwards, these directional filters are used to filter the hemodynamic feature map. For each pixel, the filter response values ​​in all directions and scales are calculated to obtain multi-scale directional response values. Assume that according to the analysis of the vascular direction in the hemodynamic feature map, the blood vessels in the vertical direction are more prominent, so the filter response value in the vertical direction is given a higher weight, for example, the weight coefficient is 1.2, and the weight coefficients in other directions are 1. The multi-scale directional response value of each pixel is weighted to obtain the weighted maximum response value and the corresponding optimal direction parameter and optimal scale parameter. For example, the maximum response value of a pixel is 0.8, the corresponding optimal direction is 90 degrees (vertical direction), and the optimal scale is 5 pixels. Generate an enhanced vascular image based on the maximum response value.

[0040] The enhanced vascular image is subjected to local maximum suppression processing to obtain a set of candidate points for the vascular centerline. For example, by comparing the response values ​​of each pixel with the eight pixels around it, if the response value of the pixel is the largest, it is marked as a candidate point for the vascular centerline. Assume that 1,000 candidate points are obtained.

[0041] An undirected graph is constructed, and the spatial position of each point in the candidate point set of the vascular centerline is used as the node coordinate in the undirected graph. A weight function for node connection is constructed based on the spatial distance between the node coordinates, the direction angle difference determined by the optimal direction parameter, and the response strength determined by the maximum response value. For example, the closer the distance between two nodes, the smaller the direction angle difference, and the larger the response value, the greater the weight between them, indicating a higher possibility of connection.

[0042] The minimum spanning tree algorithm is used to process undirected graphs. The connection priority between nodes is determined according to the value of the weight function. The dynamic weight threshold is determined according to the vascular density in the enhanced vascular image. For example, the higher the vascular density, the smaller the dynamic weight threshold is to avoid over-connection. The node connection relationship greater than the dynamic weight threshold is eliminated to obtain the initial vascular topology.

[0043] On the path corresponding to the connecting edge of the initial vascular topology, the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics are extracted from the hemodynamic feature map. For example, on a connecting edge, the blood flow velocity value of every other pixel is extracted, and the difference between the velocity values ​​of adjacent pixels is calculated as the blood flow complexity feature. And the gradient change value of each feature along the path is calculated. The feature consistency index is constructed based on the gradient change value. For example, if the gradient change value of the blood flow complexity feature on a connecting edge is very small, it is considered that the feature consistency of the connecting edge is high.

[0044] Verify the connection relationship in the initial vascular topology structure according to the feature consistency index. Identify the connection relationship with a gradient change value greater than the preset gradient threshold as an abnormal connection. For example, if the gradient change value of the blood flow complexity feature on a connection edge is greater than the preset threshold 0.5, the connection is considered to be an abnormal connection. Correct or delete the abnormal connection in combination with the hemodynamic characteristic map. For example, if an abnormal connection spans a low-speed area, delete it. Complete the reconstruction of vascular connectivity and generate vascular topology data containing node location information and connection relationship information, for example, a vascular topology data containing 100 nodes and 90 connection edges.

[0045] Existing vascular topology reconstruction methods usually rely on traditional edge detection or vascular segmentation based on statistical models. These methods are difficult to accurately extract vascular structures in complex blood flow environments, especially in areas where hemodynamic characteristics change dramatically, which can easily lead to topological structure breaks or incorrect connections. In addition, when dealing with vascular connectivity, existing methods often use fixed thresholds or simple morphological operations, and fail to fully combine blood flow characteristics for optimization, resulting in limited accuracy of topological structures. This application enhances vascular images through directional filters and reconstructs vascular topological structures in combination with the minimum spanning tree algorithm. The directional filter determines the scale range based on the vascular morphological characteristics, and generates filters of different directions by rotating the Gaussian second-order derivative filter, so that it can adapt to vascular structures of different scales and directions, and improve the contrast and clarity of vascular images. Subsequently, on the enhanced vascular image, the candidate points of the vascular centerline are extracted by the local maximum suppression method, and an undirected graph is constructed. The weight threshold is dynamically adjusted in combination with the vascular density information to ensure that the connectivity of the vascular topological structure is more consistent with the real anatomical structure. Furthermore, in the process of topological structure optimization, the present application introduces hemodynamic characteristics, combines characteristics such as blood flow vortex distribution, vascular wall stress, blood flow complexity and pressure wave propagation, calculates the gradient change value to construct a feature consistency index, and realizes accurate identification and correction of abnormal connections. Compared with existing methods, the present application not only optimizes the vascular enhancement process based on vascular morphological characteristics, but also verifies and corrects the topological structure in combination with hemodynamic characteristics, which can more accurately identify abnormal connections and improve the reliability of vascular connectivity. It significantly improves the accuracy and stability of vascular topological structure reconstruction and provides more sophisticated vascular model support for hemodynamic analysis.

[0046] In an optional embodiment, the vascular topology data is input into a recursive neural tensor network to calculate the risk value of the vascular branch node to generate a vascular lesion risk distribution map, and marking the potential embolism risk area based on the vascular lesion risk distribution map includes: Extracting feature vectors of vascular branch nodes from vascular topological structure data, calculating a topological relationship matrix between nodes, wherein the topological relationship matrix records the superior-subordinate and the same-level connection relationships of nodes, and constructing a tree-shaped computational graph based on the feature vectors and the topological relationship matrix, wherein each node in the graph contains corresponding feature vectors and topological relationship data; Inputting the tree-shaped computation graph into a recursive neural tensor network, wherein the recursive neural tensor network comprises a third-order tensor computation unit, wherein the third-order tensor computation unit performs a tensor product operation on feature vectors of adjacent nodes to obtain structural correlation data between nodes and node feature correlation strength; Based on the inter-node structural correlation data and the node feature association strength, an attention weight matrix is ​​constructed, wherein the attention weight matrix includes a feature dimension weight component and a node association weight component, and the feature vector is weighted according to the feature dimension weight component and the node association weight component to generate a weighted feature vector that integrates local structural information; The weighted feature vector is mapped using a nonlinear activation function to obtain an initial risk assessment value, and the initial risk assessment value is input into a recursive calculation module. The recursive calculation module calculates from the terminal branch node in a bottom-up manner, and fuses the feature information and risk information of the lower-level node based on the feature dimension weight component and the node association weight component, and updates the risk assessment value of the upper-level node until all nodes are traversed to generate a final risk assessment value; The final risk assessment value is normalized to generate a risk distribution map, the vascular structure is partitioned according to the risk value in the risk distribution map, and the area with a risk value higher than a set risk threshold is marked as a potential embolism risk area.

[0047] Exemplarily, first, the feature vector of the vascular branch node is extracted from the vascular topology data. The topology data can be obtained from medical imaging data (such as computed tomography angiography (Computed TomographyAngiography), magnetic resonance angiography (Magnetic Resonance Angiography)), and obtained through preprocessing steps such as image segmentation and skeleton extraction. The feature vector may include geometric features such as diameter, curvature, length, and bifurcation angle of the vascular branch node, as well as hemodynamic features such as blood flow velocity and blood flow direction. For example, the feature vector of a branch node can be expressed as [diameter: 2.5mm, curvature: 0.1, length: 5mm, bifurcation angle: 45 degrees, blood flow velocity: 10cm / s].

[0048] Next, the topological relationship matrix between the vascular branch nodes is calculated. This matrix records the superior-subordinate connection relationship and the peer connection relationship of the vascular branch nodes. For example, if node A is the parent node of node B and node C, then the values ​​of row A, column B and row A, column C in the matrix are 1, indicating a parent-child relationship; if node B and node C are peer nodes, then the values ​​of row B, column C and row C, column B in the matrix are 1, indicating a peer relationship. For other unrelated nodes, the matrix value is set to 0.

[0049] Then, a tree-like computational graph is constructed based on the eigenvector and the topological relationship matrix. Each node in the tree-like computational graph contains the corresponding eigenvector and topological relationship matrix data. The tree-like computational graph reflects the hierarchical structure of vascular branches, with the root node representing the main blood vessel and the leaf node representing the terminal vascular branch.

[0050] The constructed tree-shaped computation graph is input into the recursive neural tensor network. The network contains a third-order tensor computation unit. The third-order tensor computation unit receives the feature vectors of adjacent vascular branch nodes in the tree-shaped computation graph, performs tensor product operations on these feature vectors, and obtains the structural correlation data between nodes and the node feature association strength. For example, the tensor product of node A and its child node B is calculated to obtain a third-order tensor, which reflects the association strength between A and B in different feature dimensions.

[0051] The attention weight matrix is ​​constructed based on the structural correlation data between nodes and the node feature association strength. The matrix contains feature dimension weight components and node association weight components. For example, if the diameter feature association strength of nodes A and B is very high, then the weight corresponding to the diameter feature dimension will be larger; if the topological relationship between nodes A and B is closer, then the node association weight will be larger.

[0052] The feature vector is weighted according to the feature dimension weight component and the node association weight component to generate a weighted feature vector that integrates the local structural information. For example, the feature vector of node A is multiplied by the corresponding weight to obtain a weighted feature vector that integrates the structural information of node A and its adjacent nodes.

[0053] The weighted feature vector is mapped using a nonlinear activation function to obtain an initial risk assessment value of the vascular branch node. For example, the weighted feature vector is mapped to between 0 and 1 using a Sigmoid function to obtain an initial risk value.

[0054] The initial risk assessment value is input into the recursive calculation module. This module uses a bottom-up approach to start the calculation from the terminal branch node, fuses the feature information and risk information of the lower-level nodes based on the feature dimension weight component and the node association weight component, and updates the risk assessment value of the upper-level node until all nodes are traversed to generate the final risk assessment value. For example, the risk values ​​of nodes B and C are 0.8 and 0.6 respectively, and the risk value of the parent node A is calculated and updated according to the weight.

[0055] The final risk assessment value is normalized to generate a risk distribution map representing the probability of vascular lesions. For example, the risk values ​​of all nodes are scaled to between 0 and 1 to generate a risk distribution map.

[0056] The vascular structure is partitioned according to the risk value in the risk distribution map, and the area with a risk value higher than the set risk threshold is marked as a potential embolism risk area. For example, the area with a risk value higher than 0.8 is marked as a potential embolism risk area.

[0057] Existing vascular lesion risk assessment methods usually rely on rule-based statistical models or traditional machine learning methods, which make it difficult to fully utilize the vascular topology and its internal hemodynamic characteristics, resulting in inaccurate risk assessment of complex vascular branch areas. In addition, existing methods are mostly based on separate feature indicators for risk assessment, and fail to effectively integrate the correlation between multiple features, which makes it limited in identifying potential embolic risk areas. This application calculates vascular topology data through a recursive neural tensor network to obtain the risk value of vascular branch nodes and generate a vascular lesion risk distribution map. A tree-shaped calculation graph is constructed using vascular topology data, so that the hierarchical relationship of vascular branches can be clearly expressed, and the geometric morphological characteristics and hemodynamic characteristics of the blood vessels are combined to more comprehensively characterize vascular structure information. A third-order tensor calculation unit is used to perform tensor product operations on the feature vectors of adjacent vascular branch nodes to obtain the structural correlation data between nodes and the node feature correlation strength. Compared with traditional methods, this method can effectively capture the interactive relationship between different features and improve the accuracy of risk assessment. During the risk calculation process, this application constructs an attention weight matrix to simultaneously consider the weight components of the feature dimensions and the node-associated weight components, so that the risk assessment results can dynamically adjust the importance of features and enhance the ability to identify key risk factors. In addition, through the recursive calculation module, a bottom-up approach is adopted to fuse the feature information and risk assessment values ​​of the lower-level nodes step by step to ensure that the risk assessment of the vascular trunk fully considers the influence of all branches, thereby improving the rationality of the overall risk assessment. Finally, a risk distribution map is generated through normalization processing, and high-risk areas are accurately marked. Compared with existing methods, this application can more accurately identify potential embolism risk areas, improve the early diagnosis of vascular lesions, and provide more reliable support for clinical decision-making.

[0058] In an optional embodiment, a multispectral imaging device is used to obtain tissue perfusion parameters of a potential embolic risk area, and electroencephalogram signals of the potential embolic risk area are collected to obtain brain tissue function parameters. The tissue perfusion parameters and brain tissue function parameters are adaptively threshold segmented to obtain an ischemic penumbra area, including: A multispectral imaging device is used to perform continuous wavelength scanning within a preset wavelength range through a liquid crystal tunable filter, and multiple frames of images are accumulated and collected at each wavelength point, and the average value is taken to obtain a spectral data cube; A tissue optical parameter inversion model is constructed according to the incident light intensity, the transmitted light intensity and the tissue thickness, the spectral data cube is input into the tissue optical parameter inversion model, the oxygenated hemoglobin concentration and the deoxygenated hemoglobin concentration are calculated by the least square fitting method, the tissue blood oxygen saturation is calculated by the ratio of the oxygenated hemoglobin concentration to the sum of the oxygenated hemoglobin concentration and the deoxygenated hemoglobin concentration, and the tissue reflectance ratio is calculated at a preset characteristic wavelength to obtain the tissue perfusion index; The EEG signal of the potential embolism risk area is collected and band-pass filtered, and the filtered EEG signal is decomposed into a plurality of preset frequency bands by using a wavelet packet decomposition method, and the ratio of the energy of each preset frequency band to the total energy is calculated to obtain a relative power spectrum, and the filtered EEG signal is converted into a symbol sequence, and then the probability distribution of different arrangement patterns is statistically obtained to obtain an arrangement entropy value, and the relative power spectrum and the arrangement entropy value are combined in a preset order to obtain brain tissue function parameters; The tissue perfusion parameters and brain tissue function parameters are combined into a feature vector, the Euclidean distance from each sample point in the feature vector to each cluster center is calculated, the membership of each sample point to each cluster center is calculated based on the Euclidean distance, the cluster center is updated according to the membership, and the iteration is repeated until the difference between the objective functions of two adjacent iterations is less than a preset convergence threshold to obtain the final cluster center; The feature vector is classified based on the final cluster center to obtain the spatial distribution of ischemic area, penumbra area and normal area, a set of pixel points in the penumbra area is extracted for edge detection to obtain a closed boundary contour, and the area within the closed boundary contour is marked as the ischemic penumbra area.

[0059] Exemplarily, first, a multispectral imaging device is used to obtain tissue perfusion parameters in the potential embolism risk area. For example, a multispectral camera equipped with a specific band filter is used to collect a multispectral image of the area. By analyzing the light intensity information of different bands, parameters reflecting the tissue blood perfusion situation, such as hemoglobin concentration, blood oxygen saturation, etc., can be obtained. Assuming that images of the red light band and the near-infrared band are collected, a parameter reflecting the tissue blood oxygen saturation can be obtained by calculating the light intensity ratio of the two bands.

[0060] Next, a hyperspectral imaging system is used to obtain reflectance spectral data of potential embolic risk areas. The hyperspectral imaging system includes a xenon lamp light source and a liquid crystal tunable filter. The liquid crystal tunable filter performs continuous wavelength scanning at a wavelength interval of 1 nm in the wavelength range of 400nm to 700nm. After accumulating 10 frames of images at each wavelength point, the average value is taken to obtain a spectral data cube. For example, at a pixel point, a spectral curve containing 400 to 700 wavelength points can be obtained.

[0061] Then, according to the incident light intensity, transmitted light intensity and tissue thickness, a tissue optical parameter inversion model is constructed. The acquired spectral data cube is input into the model, and the concentration of oxyhemoglobin and deoxyhemoglobin is calculated using the least squares fitting method. For example, assuming that the tissue thickness is 1 mm, the concentrations of oxyhemoglobin and deoxyhemoglobin can be calculated by measuring the incident light intensity and the transmitted light intensity, and combining the optical properties of the tissue. The tissue blood oxygen saturation is calculated by calculating the ratio of the oxyhemoglobin concentration to the sum of the oxyhemoglobin concentration and the deoxyhemoglobin concentration. The tissue perfusion index is obtained by calculating the tissue reflectance ratio at two characteristic wavelengths of 570 nm and 610 nm. For example, assuming that the reflectance at 570 nm is 0.6 and the reflectance at 610 nm is 0.5, the tissue perfusion index is 1.2.

[0062] At the same time, the EEG signals of the potential embolic risk area are collected. The EEG signals are band-pass filtered in the frequency range of 1-30Hz to obtain the filtered EEG signals. The wavelet packet decomposition method is used to decompose the filtered EEG signals into four frequency bands: delta (0.5-4Hz), theta (4-8Hz), alpha (8-13Hz) and beta (13-30Hz). The relative power spectrum is calculated by calculating the ratio of the energy of each frequency band to the total energy. For example, the relative power spectrum of the delta band is 0.2, the theta band is 0.3, the alpha band is 0.3, and the beta band is 0.2. After the filtered EEG signals are converted into symbol sequences, the probability distribution of different arrangement patterns is statistically obtained to obtain the arrangement entropy value. For example, the calculated arrangement entropy value is 0.8. The relative power spectrum and the arrangement entropy value are combined in a preset order to obtain the brain tissue function parameters of the potential embolic risk area.

[0063] The tissue perfusion parameters and brain tissue function parameters are combined into a feature vector. An adaptive threshold segmentation model based on fuzzy clustering is constructed. The Euclidean distance from each sample point in the feature vector to each cluster center is calculated. The membership of each sample point to each cluster center is calculated based on the Euclidean distance. The cluster center is updated according to the membership, and the iteration is repeated until the difference between the objective functions of two adjacent iterations is less than the convergence threshold of 0.001 to obtain the final cluster center.

[0064] Based on the final cluster center, the feature vectors are classified to obtain the spatial distribution of the ischemic area, the penumbra area, and the normal area. A set of pixel points in the penumbra area is extracted. Edge detection is performed on the pixel point set to obtain a closed boundary contour of the penumbra area. The area within the closed boundary contour is marked as the ischemic penumbra area. For example, the boundary of the penumbra area is extracted by the Canny edge detection algorithm, and the pixels within the boundary are marked as the ischemic penumbra area.

[0065] In this embodiment, through the comprehensive analysis of multispectral imaging and EEG signals, tissue perfusion parameters and brain tissue function parameters can be obtained, providing key data for the subsequent identification of ischemic areas and penumbra areas. The hyperspectral imaging system calculates the blood oxygen saturation and perfusion index of the tissue by accurately measuring the optical parameters of the tissue, ensuring a comprehensive assessment of the tissue perfusion situation. The EEG signal extracts the band energy and permutation entropy through wavelet packet decomposition, which can deeply understand the functional state of the brain tissue and assist in identifying the embolic risk area. The adaptive threshold segmentation model combined with the fuzzy clustering method effectively realizes the fusion of blood perfusion and brain function information, and accurately divides the ischemic area, penumbra area and normal area. Through edge detection technology, the boundary of the penumbra area is further clarified, and the positioning accuracy of the ischemic penumbra area is improved, providing a reliable embolic risk assessment and ischemic area identification tool for clinical use, supporting personalized treatment decisions.

[0066] Figure 2 : is a comparison chart of the accuracy of vascular connectivity reconstruction in the embodiment of the present invention, such as Figure 2 As shown in the figure, the reconstruction accuracy of this technical solution is compared with that of the regional growing method and the Dijkstra shortest path method at different branch levels. The horizontal axis represents the vascular branch level (level 1-4), and the vertical axis represents the reconstruction accuracy (60%-100%). This technical solution achieves the highest accuracy of 98.5% at the first-level branch. As the branch level increases, the accuracy decreases but remains at a high level, reaching 91.6% at the fourth-level branch. In contrast, the accuracy of the regional growing method is 90% at the first-level branch, which drops to 78% at the fourth-level branch; the Dijkstra shortest path method performs the worst, with an accuracy of 85% at the first-level branch and 72% at the fourth-level branch. The data show that this technical solution significantly improves the reconstruction accuracy of multi-level vascular structures by introducing hemodynamic characteristics and feature consistency verification mechanisms, especially when dealing with high-level branches. The specific accuracy values ​​of each level are clearly marked in the figure, highlighting the superiority of this technical solution in vascular connectivity reconstruction.

[0067] Figure 3 This is a diagram showing the effect of vascular risk distribution according to an embodiment of the present invention, such as Figure 3As shown in the figure, the results of the vascular network risk distribution analysis are shown. In the axial position of the main blood vessel, three representative risk areas are clearly identified: the risk value of the Risk-A area on the left reaches 0.95, located at 150mm from the starting section of the blood vessel, showing a typical Gaussian distribution feature; the risk value of the Risk-B area in the middle is 0.88, located at 400mm, and the risk distribution is more concentrated; the risk value of the Risk-C area on the right is 0.82, located at 650mm, showing a gradual distribution feature. Through 50,000 Monte Carlo sampling analyses, the overall risk distribution is obtained to be a log-normal distribution with a mean of 0.82±0.03. The system detection rate reaches 96.8%, the false alarm rate is only 2.3%, and the calculation time is controlled within 55ms, meeting the requirements of real-time analysis. Sufficient spatial separation is maintained between the risk areas to ensure the accuracy and reliability of risk assessment. The risk distribution curve fully reflects the risk change trend along the axial direction of the blood vessel, providing an accurate spatial positioning basis for subsequent clinical intervention. The analysis results fully verify the superiority of this scheme in vascular network risk assessment.

[0068] Figure 4 FIG. 1 is a schematic diagram of a structure of an early intelligent screening and warning system for acute stroke based on deep learning according to an embodiment of the present invention. Figure 4 As shown, the system comprises: The first unit is used to trigger a scanning sequence to acquire multi-phase angiography images within a cardiac cycle through a pulsed dynamic imaging scheme, perform hemodynamic analysis on the multi-phase angiography images to obtain blood flow velocity vector field data, construct a blood flow feature tensor according to the blood flow velocity vector field data, input the blood flow feature tensor into a dual-flow feature extraction network, wherein the vascular structure feature is extracted through a spatial feature branch, the blood flow time series feature is extracted through a time series feature branch, and the vascular structure feature is fused with the blood flow time series feature to generate a hemodynamic feature map; The second unit is used to process the hemodynamic characteristic map through a directional filter to obtain an enhanced vascular image, reconstruct the vascular topology data using a minimum spanning tree algorithm, input the vascular topology data into a recursive neural tensor network to calculate the risk value of the vascular branch node to generate a vascular lesion risk distribution map, and mark the potential embolism risk area based on the vascular lesion risk distribution map; The third unit is used to obtain tissue perfusion parameters of potential embolic risk areas using multispectral imaging equipment, collect electroencephalogram signals of potential embolic risk areas to obtain brain tissue function parameters, perform adaptive threshold segmentation on tissue perfusion parameters and brain tissue function parameters to obtain the ischemic penumbra area, calculate the area value of the ischemic penumbra area, input the area value of the ischemic penumbra area, tissue perfusion parameters and brain tissue function parameters into a hierarchical decision model to calculate the risk level value, and generate a clinical intervention plan based on the risk level value.

[0069] According to a third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0070] According to a fourth aspect of the embodiments of the present invention, A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0071] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An early intelligent screening and warning method for acute stroke based on deep learning, characterized in that: include: A pulsed dynamic imaging scheme is used to trigger a scanning sequence within the cardiac cycle to acquire multi-phase angiographic images, and a hemodynamic analysis is performed on the multi-phase angiographic images to obtain blood flow velocity vector field data. A blood flow feature tensor is constructed based on the blood flow velocity vector field data, and the blood flow feature tensor is input into a dual-flow feature extraction network, wherein the vascular structure features are extracted through a spatial feature branch, and the blood flow time series features are extracted through a time series feature branch, and the vascular structure features and the blood flow time series features are fused to generate a hemodynamic feature map; The hemodynamic characteristic map is processed by a directional filter to obtain an enhanced vascular image, and the vascular topology data is reconstructed using a minimum spanning tree algorithm. The vascular topology data is input into a recursive neural tensor network to calculate the risk value of vascular branch nodes to generate a vascular lesion risk distribution map, and potential embolism risk areas are marked based on the vascular lesion risk distribution map. A multispectral imaging device is used to obtain tissue perfusion parameters in potential embolism risk areas, and electroencephalogram signals in potential embolism risk areas are collected to obtain brain tissue function parameters. The tissue perfusion parameters and brain tissue function parameters are adaptively thresholded to obtain the ischemic penumbra area, and the area value of the ischemic penumbra area is calculated. The area value of the ischemic penumbra area, tissue perfusion parameters, and brain tissue function parameters are input into a hierarchical decision model to calculate the risk level value, and a clinical intervention plan is generated according to the risk level value.

2. The method according to claim 1, characterized in that The pulsed dynamic imaging scheme is used to trigger the scanning sequence to acquire multi-phase angiographic images during the cardiac cycle, and the multi-phase angiographic images are subjected to hemodynamic analysis to obtain blood flow velocity vector field data. The blood flow characteristic tensor is constructed based on the blood flow velocity vector field data, including: The patient's electrocardiogram (ECG) signal is collected to construct a pulsed dynamic imaging scheme, wherein the pulsed dynamic imaging scheme uses a dual-domain sparse sampling model in the frequency domain space and the time domain for undersampling, and multiple trigger windows are constructed within the cardiac cycle based on the patient's ECG signal. A spiral angle incremental ray sampling strategy is used within the multiple trigger windows to trigger a scanning sequence, thereby obtaining non-uniform grid data for multi-phase angiography; Iteratively reconstructing the non-uniform grid data, optimizing the data consistency in the spatial domain and the temporal domain in each iteration, and finally obtaining a multi-phase angiography image with high temporal and spatial resolution; Inputting the multi-phase angiography image into a multi-scale hemodynamic feature extraction network, the multi-scale hemodynamic feature extraction network includes an adaptive deformation convolution layer, a spatial attention module, and a cyclic feature extraction unit, wherein the adaptive deformation convolution layer extracts local blood flow features to obtain vascular feature data, the spatial attention module determines the vascular branch connection relationship based on the vascular feature data to obtain vascular network data, and the cyclic feature extraction unit performs time series analysis on the vascular network data based on pressure wave propagation constraints, and finally obtains blood flow velocity vector field data; Phase enhancement technology is used to calculate the phase difference information between adjacent phase images that characterizes the magnitude and direction of blood flow velocity. Based on the phase difference information, nonlinear correction is performed on the blood flow velocity vector field data. According to the corrected blood flow velocity vector field data, a blood flow characteristic tensor including blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics is constructed.

3. The method according to claim 1, characterized in that The blood flow feature tensor is input into the dual-flow feature extraction network, wherein the vascular structure features are extracted through the spatial feature branch, and the blood flow timing features are extracted through the timing feature branch. The vascular structure features and the blood flow timing features are fused to generate the hemodynamic feature map, including: Inputting the blood flow feature tensor into a dual-flow feature extraction network, the dual-flow feature extraction network includes a spatial feature branch and a temporal feature branch, the blood flow feature tensor includes blood flow vortex distribution features, blood vessel wall stress distribution features, blood flow complexity features and pressure wave propagation features; The spatial feature branch generates a convolution kernel deformation parameter based on the vascular morphology information in the blood flow feature tensor, performs multi-scale convolution operations on the deformable convolution kernel group and the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics to obtain a local feature atlas, and performs channel dimension weighting on the local feature atlas to obtain vascular structure characteristics; The time series feature branch performs phase analysis on the blood flow vortex distribution characteristics, vascular wall stress distribution characteristics, blood flow complexity characteristics and pressure wave propagation characteristics in the blood flow feature tensor, extracts the instantaneous phase and instantaneous amplitude of each characteristic component through Hilbert transform, calculates the phase difference between time adjacent windows to obtain a phase change curve group, constructs a phase consistency matrix based on the phase change curve group, selectively enhances the instantaneous amplitude and reconstructs it to obtain the blood flow time series characteristics; The mutual information between the vascular structure features and the blood flow timing features is calculated to obtain a feature association matrix, a bidirectional attention weight is constructed based on the feature association matrix, the vascular structure features and the blood flow timing features are adaptively enhanced respectively, and the enhanced features are fused through a jump connection structure to obtain a hemodynamic feature map.

4. The method according to claim 1, characterized in that: The hemodynamic characteristic map is processed by a directional filter to obtain an enhanced vascular image, and the minimum spanning tree algorithm is used to reconstruct the vascular topology data including: Extracting blood vessel morphological features in the hemodynamic characteristic map, determining the scale range of a Gaussian second-order derivative filter based on the blood vessel morphological features, and rotating the Gaussian second-order derivative filter within a preset angle range to generate a directional filter; Filtering the hemodynamic characteristic map using the directional filter to obtain multi-scale directional response values, determining weighting coefficients according to the direction of blood vessels in the hemodynamic characteristic map, weighting the multi-scale directional response values, obtaining the weighted maximum response value and the corresponding optimal direction parameters and optimal scale parameters, and generating an enhanced blood vessel image; Performing local maximum suppression processing on the enhanced vascular image, obtaining a set of candidate points for the vascular centerline, constructing an undirected graph, processing the undirected graph using a minimum spanning tree algorithm, determining a dynamic weight threshold according to the vascular density in the enhanced vascular image, removing node connections greater than the dynamic weight threshold, and obtaining an initial vascular topology structure; On the path corresponding to the connection edge of the initial vascular topological structure, blood flow vortex distribution, vascular wall stress distribution, blood flow complexity and pressure wave propagation characteristics are extracted, and the gradient change value is calculated to construct a feature consistency index; The connection relationship in the initial vascular topology structure is verified according to the feature consistency index, and the connection relationship with a gradient change value greater than a preset gradient threshold is identified as abnormal and corrected or deleted, so as to complete the vascular connectivity reconstruction and generate vascular topology structure data.

5. The method according to claim 1, characterized in that The vascular topology data is input into the recursive neural tensor network to calculate the risk value of the vascular branch node to generate a vascular lesion risk distribution map. Based on the vascular lesion risk distribution map, the potential embolism risk areas are marked, including: Extracting feature vectors of vascular branch nodes from vascular topological structure data, calculating a topological relationship matrix between nodes, wherein the topological relationship matrix records the superior-subordinate and the same-level connection relationships of nodes, and constructing a tree-shaped computational graph based on the feature vectors and the topological relationship matrix, wherein each node in the graph contains corresponding feature vectors and topological relationship data; Inputting the tree-shaped computation graph into a recursive neural tensor network, wherein the recursive neural tensor network comprises a third-order tensor computation unit, wherein the third-order tensor computation unit performs a tensor product operation on feature vectors of adjacent nodes to obtain structural correlation data between nodes and node feature correlation strength; Based on the inter-node structural correlation data and the node feature association strength, an attention weight matrix is ​​constructed, wherein the attention weight matrix includes a feature dimension weight component and a node association weight component, and the feature vector is weighted according to the feature dimension weight component and the node association weight component to generate a weighted feature vector that integrates local structural information; The weighted feature vector is mapped using a nonlinear activation function to obtain an initial risk assessment value, and the initial risk assessment value is input into a recursive calculation module. The recursive calculation module calculates from the terminal branch node in a bottom-up manner, and fuses the feature information and risk information of the lower-level node based on the feature dimension weight component and the node association weight component, and updates the risk assessment value of the upper-level node until all nodes are traversed to generate a final risk assessment value; The final risk assessment value is normalized to generate a risk distribution map, the vascular structure is partitioned according to the risk value in the risk distribution map, and the area with a risk value higher than a set risk threshold is marked as a potential embolism risk area.

6. The method according to claim 1, characterized in that Multispectral imaging equipment is used to obtain tissue perfusion parameters in potential embolic risk areas, and electroencephalogram signals in potential embolic risk areas are collected to obtain brain tissue function parameters. The tissue perfusion parameters and brain tissue function parameters are segmented by adaptive thresholds to obtain the ischemic penumbra area, including: A multispectral imaging device is used to perform continuous wavelength scanning within a preset wavelength range through a liquid crystal tunable filter, and multiple frames of images are accumulated and collected at each wavelength point, and the average value is taken to obtain a spectral data cube; A tissue optical parameter inversion model is constructed according to the incident light intensity, the transmitted light intensity and the tissue thickness, the spectral data cube is input into the tissue optical parameter inversion model, the oxygenated hemoglobin concentration and the deoxygenated hemoglobin concentration are calculated by the least square fitting method, the tissue blood oxygen saturation is calculated by the ratio of the oxygenated hemoglobin concentration to the sum of the oxygenated hemoglobin concentration and the deoxygenated hemoglobin concentration, and the tissue reflectance ratio is calculated at a preset characteristic wavelength to obtain the tissue perfusion index; The EEG signal of the potential embolism risk area is collected and band-pass filtered, and the filtered EEG signal is decomposed into a plurality of preset frequency bands by using a wavelet packet decomposition method, and the ratio of the energy of each preset frequency band to the total energy is calculated to obtain a relative power spectrum, and the filtered EEG signal is converted into a symbol sequence, and then the probability distribution of different arrangement patterns is statistically obtained to obtain an arrangement entropy value, and the relative power spectrum and the arrangement entropy value are combined in a preset order to obtain brain tissue function parameters; The tissue perfusion parameters and brain tissue function parameters are combined into a feature vector, the Euclidean distance from each sample point in the feature vector to each cluster center is calculated, the membership of each sample point to each cluster center is calculated based on the Euclidean distance, the cluster center is updated according to the membership, and the iteration is repeated until the difference between the objective functions of two adjacent iterations is less than a preset convergence threshold to obtain the final cluster center; The feature vector is classified based on the final cluster center to obtain the spatial distribution of ischemic area, penumbra area and normal area, a set of pixel points in the penumbra area is extracted for edge detection to obtain a closed boundary contour, and the area within the closed boundary contour is marked as the ischemic penumbra area.

7. The method according to claim 1, characterized in that The area value of the ischemic penumbra region is calculated, and the area value of the ischemic penumbra region, tissue perfusion parameters, and brain tissue function parameters are input into the hierarchical decision model to calculate the risk level value. The clinical intervention plan generated according to the risk level value includes: Marking the connected domains of the ischemic penumbra area and constructing a neuron connection map, calculating the topological relationship matrix between neurons, counting the total number of pixels in the ischemic penumbra area, and converting the total number of pixels into an area value of the ischemic penumbra area according to a spatial resolution parameter; A hierarchical decision model of the spatiotemporal attention mechanism is constructed, including a spatial attention module and a temporal attention module; the spatial attention module uses a wavelet multi-resolution analysis method to decompose the tissue perfusion parameters and the topological relationship matrix into spatial feature matrices of multiple scales, and extracts the dynamic change characteristics of the tissue perfusion parameters through a recursive neural network to calculate the spatial weight coefficient; the temporal attention module uses a nonlinear time series decomposition method to decompose the brain tissue function parameters into feature sequences of different time scales, and extracts the time series change characteristics through an attention autoencoder to calculate the time weight coefficient; A multimodal risk assessment feature set is constructed according to the area value of the ischemic penumbra region, the weighted tissue perfusion parameter, and the weighted brain tissue function parameter, and the fused feature vector is input into a graph convolutional neural network together with a neuron connection map to obtain a fused feature vector, and the fused feature vector is input into a residual network to obtain a risk level value; The risk evolution trend is predicted based on the risk level value, and clinical intervention rules are selected in combination with historical intervention effect data to generate a clinical intervention decision tree. The decision tree nodes are prioritized based on the risk evolution trend, and intervention effect and resource consumption are used as evaluation indicators to obtain a clinical intervention plan with priority.

8. An early intelligent screening and warning system for acute stroke based on deep learning, used to implement the method described in any one of claims 1 to 7, characterized in that: include: The first unit is used to trigger a scanning sequence to acquire multi-phase angiography images within a cardiac cycle through a pulsed dynamic imaging scheme, perform hemodynamic analysis on the multi-phase angiography images to obtain blood flow velocity vector field data, construct a blood flow feature tensor according to the blood flow velocity vector field data, input the blood flow feature tensor into a dual-flow feature extraction network, wherein the vascular structure feature is extracted through a spatial feature branch, the blood flow time series feature is extracted through a time series feature branch, and the vascular structure feature is fused with the blood flow time series feature to generate a hemodynamic feature map; The second unit is used to process the hemodynamic characteristic map through a directional filter to obtain an enhanced vascular image, reconstruct the vascular topology data using a minimum spanning tree algorithm, input the vascular topology data into a recursive neural tensor network to calculate the risk value of the vascular branch node to generate a vascular lesion risk distribution map, and mark the potential embolism risk area based on the vascular lesion risk distribution map; The third unit is used to obtain tissue perfusion parameters of potential embolic risk areas using multispectral imaging equipment, collect electroencephalogram signals of potential embolic risk areas to obtain brain tissue function parameters, perform adaptive threshold segmentation on tissue perfusion parameters and brain tissue function parameters to obtain the ischemic penumbra area, calculate the area value of the ischemic penumbra area, input the area value of the ischemic penumbra area, tissue perfusion parameters and brain tissue function parameters into a hierarchical decision model to calculate the risk level value, and generate a clinical intervention plan based on the risk level value.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method described in any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images