Method and device for establishing and applying whole-spine morphological model based on medical image

Abstract

The invention provides a method and a device for establishing and applying a whole-spine morphological model based on a medical image, and relates to the technical field of medical images. The method comprises the following steps: acquiring morphological characteristics and medical image data of the whole spine of a patient, and constructing a whole spine three-dimensional morphological architecture; generating a full-spine morphological model based on the full-spine three-dimensional morphological architecture; obtaining a personalized correction demand of a patient, and generating morphological difference information based on the personalized correction demand; generating a morphological matching chain according to the morphological difference information and the whole spine morphological model; generating a personalized correction model of the patient according to the morphological matching chain; generating a resource allocation strategy of the patient based on the personalized correction model and the whole spine morphological model; wherein the resource allocation strategy is used for allocating three-dimensional modeling nodes and data processing resources for the personalized correction requirements. According to the method, the precise modeling and personalized correction of the whole spine can be realized, and the scientificity of spinal disease diagnosis and treatment and the resource utilization efficiency are improved.

Description

Technical Field

[0001] This invention relates to the field of medical imaging technology, and in particular to a method and apparatus for establishing and applying a full spinal morphological model based on medical images. Background Technology

[0002] In the diagnosis, treatment, and research of spinal diseases, precise analysis of the entire spinal morphology is of paramount importance. Traditional methods of spinal morphological assessment mainly rely on single medical imaging data or simple surface measurements, which have significant limitations. For example, X-ray or CT images alone cannot fully reflect the relationship between the surface morphological features of the spine and its internal structures, while relying solely on surface parameters such as height and weight cannot deeply integrate the anatomical information contained in medical images. This results in an incomplete and inaccurate assessment of spinal morphology, making it difficult to meet the needs of personalized diagnosis and treatment.

[0003] With the continuous development of medical imaging technology, acquiring multimodal imaging data has become more convenient. However, effectively integrating this data and constructing a model that accurately reflects the morphological characteristics of the entire spine remains a challenge. Current technologies lack a method for deeply correlating and fusing surface morphological parameters with medical imaging data, making it impossible to model the morphology of the entire spine from multiple angles and levels. Furthermore, when addressing patients' personalized correction needs, traditional methods struggle to quickly and accurately generate matching correction models, and lack scientific and rational strategies for resource allocation, resulting in inefficient correction plan development and suboptimal resource utilization.

[0004] In the treatment of spinal diseases, the development of personalized correction plans requires full consideration of the individual morphological differences of patients. However, current technologies lack a systematic approach to processing and utilizing morphological differences, failing to generate accurate morphological matching chains based on a full spinal morphological model, thus affecting the accuracy and effectiveness of personalized correction models. Furthermore, ensuring data integrity, spatial consistency, and processing efficiency during data processing and model construction are also pressing issues that current technologies need to address. Summary of the Invention

[0005] To address the technical problems mentioned in the background section, this invention provides a method and apparatus for establishing and applying a full spine morphological model based on medical imaging. The technical solution is as follows:

[0006] On the one hand, a method for establishing and applying a whole-spine morphological model based on medical imaging is provided. This method is implemented by a whole-spine morphological model establishment and application device, and includes:

[0007] S1. Obtain the morphological features and medical imaging data of the patient's entire spine, construct a surface morphological parameter layer based on the morphological features, construct a medical imaging data layer based on the medical imaging data, and construct a three-dimensional morphological architecture of the entire spine based on the surface morphological parameter layer and the medical imaging data layer.

[0008] S2. Generate a full spine morphological model based on the three-dimensional morphological architecture of the entire spine.

[0009] S3. Obtain the patient's personalized correction needs and generate morphological difference information based on the personalized correction needs; generate a morphological matching chain based on the morphological difference information and the whole spine morphological model.

[0010] S4. Generate a personalized correction model for the patient based on the morphological matching chain.

[0011] S5. Generate a resource allocation strategy for patients based on a personalized correction model and a full spinal morphology model; wherein, the resource allocation strategy is used to allocate 3D modeling nodes and data processing resources for personalized correction needs.

[0012] Optionally, the generation of a full-spine morphological model based on a full-spine three-dimensional morphological architecture in S2 includes:

[0013] S21. Based on the three-dimensional morphological architecture of the whole spine, determine the association status of each parameter node in the body surface morphology parameter layer and the spatial distribution status of each image data node in the medical image data layer.

[0014] S22. Obtain the parameter distribution information of the body surface morphology parameter layer and the image topology information of the medical image data layer based on the association status of each parameter node and the spatial distribution status of each image data node.

[0015] S23. Obtain the attribute information of each parameter node and the attribute information of each image data node based on the parameter distribution information and image topology information.

[0016] S24. Obtain the data resource information of each parameter node and the data resource information of each image data node.

[0017] S25. Construct a body surface parameter model based on the attribute information and data resource information of each parameter node, and construct an image data model based on the attribute information and data resource information of each image data node.

[0018] S26. Construct a full spinal morphological model based on the body surface parameter model and the imaging data model.

[0019] Optionally, S3 generates morphological matching chains based on morphological difference information and a full spine morphological model, including:

[0020] S31. Generate a target parameter set based on morphological difference information and the surface morphological parameter layer of the whole spine morphological model; wherein, the target parameter set includes multiple target morphological nodes.

[0021] S32. Generate a morphological component sequence based on the target parameter set and personalized correction needs; wherein, the morphological component sequence includes multiple morphological components.

[0022] S33. Generate a morphological matching chain based on the association between each target morphological node in the target parameter set and the morphological component sequence.

[0023] Optionally, in S31, a set of target parameters is generated based on morphological difference information and the surface morphological parameter layer of the whole spine morphological model, including:

[0024] S311. Based on morphological difference information, multiple candidate morphological nodes are selected from the surface morphological parameter layer of the whole spine morphological model.

[0025] S312. Generate multiple candidate parameter sets based on multiple candidate morphological nodes.

[0026] S313. Perform morphological evaluation and resource evaluation on each candidate parameter set to obtain the evaluation result of each candidate parameter set; wherein, the evaluation result includes scores of multiple evaluation dimensions; the multiple evaluation dimensions include parameter matching degree dimension, data integrity dimension, spatial consistency dimension and resource consumption dimension.

[0027] S314. Determine the weighting coefficient for each evaluation dimension based on individualized correction needs.

[0028] S315. The evaluation results of each candidate parameter set are weighted based on the weight coefficients to obtain the comprehensive score of each candidate parameter set.

[0029] S316. Based on the comprehensive score of each candidate parameter set, select the target parameter set from multiple candidate parameter sets.

[0030] Optionally, S4 generates a personalized correction model for the patient based on a morphological matching chain, including:

[0031] S41. Determine the source data node set and target data node set based on personalized correction needs.

[0032] S42. Determine the data mapping link for personalized correction needs based on the source data node set and the target data node set.

[0033] S43. Determine the dependencies between morphological components in the morphological matching chain based on the data mapping link.

[0034] S44. Construct a correlation model between morphological components based on the dependencies between them.

[0035] S45. Determine the data embedding information of the morphological matching chain based on the association model between each morphological component.

[0036] S46. Generate a mapping vector based on data embedding information; wherein the mapping vector includes a component embedding vector, a data mapping vector, and a correction receiving vector.

[0037] S47. Generate a personalized correction model for the patient based on the mapping vector.

[0038] Optionally, the resource allocation strategy for generating patients in S5 based on personalized correction models and whole-spine morphology models includes:

[0039] S51. Extract morphological features and data requirements of personalized correction needs from the personalized correction model.

[0040] S52. Generate corrective constraints based on morphological feature information and data requirement information; wherein, the corrective constraints include source node constraints, target node constraints, component embedding constraints, processing time constraints, data consistency constraints, node resource constraints, and mapping bandwidth constraints of the morphological matching chain.

[0041] S53. Generate an optimization model based on the correction constraints; wherein the optimization model includes a first optimization model and a second optimization model; the first optimization model is used to minimize the allocation deviation of data processing resources; the second optimization model is used to minimize the migration cost of the morphological matching chain.

[0042] S54. Solve the optimization model to obtain the solution results; generate a resource allocation strategy for the patient based on the solution results and the full spine morphology model.

[0043] Optionally, S1 also includes:

[0044] The medical image data layer is reconstructed in three dimensions to generate a full spine contour model; the full spine contour model is used to generate a full spine morphological model.

[0045] The generation of the full spine contour model includes:

[0046] Multimodal data fusion was performed on the whole spine MRI image sequence in the medical imaging data layer to obtain fused three-dimensional image data.

[0047] The fused 3D image data is subjected to noise filtering and resolution enhancement to obtain optimized image data.

[0048] Based on the optimized image data, a full spine contour model is generated using an inter-slice interpolation algorithm.

[0049] On the other hand, a device for establishing and applying a whole-spine morphological model based on medical images is provided. This device is applied to the method of establishing and applying a whole-spine morphological model based on medical images. The device includes:

[0050] The whole spine three-dimensional morphological architecture construction module is used to acquire the morphological features and medical imaging data of the patient's whole spine. Based on the morphological features, a body surface morphological parameter layer is constructed, and based on the medical imaging data, a medical imaging data layer is constructed. Based on the body surface morphological parameter layer and the medical imaging data layer, the whole spine three-dimensional morphological architecture is constructed.

[0051] The whole spine morphology model generation module is used to generate a whole spine morphology model based on the three-dimensional morphological architecture of the whole spine.

[0052] The morphological matching chain generation module is used to obtain the patient's personalized correction needs and generate morphological difference information based on the personalized correction needs; and to generate a morphological matching chain based on the morphological difference information and the whole spinal morphological model.

[0053] The personalized correction model generation module is used to generate personalized correction models for patients based on morphological matching chains.

[0054] The output module is used to generate a resource allocation strategy for patients based on the personalized correction model and the full spinal morphology model; the resource allocation strategy is used to allocate 3D modeling nodes and data processing resources for personalized correction needs.

[0055] On the other hand, a device for establishing and applying a whole spine morphological model is provided. The device includes: a processor; a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any one of the methods described above for establishing and applying a whole spine morphological model based on medical images is implemented.

[0056] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, the at least one instruction being loaded and executed by a processor to implement any of the methods described above for establishing and applying a whole spine morphological model based on medical images.

[0057] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0058] This invention constructs a three-dimensional morphological architecture of the entire spine, comprising a layer of surface morphological parameters and a layer of medical imaging data. By matching these two layers using a data association algorithm, deep fusion of surface parameters and medical imaging data is achieved, enabling a comprehensive and accurate reflection of the morphological characteristics of the entire spine from multiple dimensions. This multi-dimensional modeling approach overcomes the limitations of traditional single-source data, providing more comprehensive and accurate foundational data for the diagnosis and treatment of spinal diseases.

[0059] In generating the full spine morphological model, by determining the association state of parameter nodes and the spatial distribution state of image data nodes, parameter distribution information and image topology information are obtained. Combined with attribute information and data resource information, this enables the model to accurately describe the characteristics of each parameter node and image data node and their interrelationships. This refined modeling process improves the accuracy and reliability of the model, providing a solid foundation for subsequent personalized correction needs analysis and resource allocation.

[0060] In response to patients' personalized correction needs, a target parameter set is generated based on morphological difference information. This is further followed by the generation of morphological component sequences and morphological matching chains. This approach accurately captures the individual morphological differences of each patient, providing a clear direction and basis for the generation of personalized correction models. By conducting multi-dimensional evaluation and weighted calculations on the candidate parameter set, the scientific validity and rationality of the target parameter set are ensured, improving the matching degree between the correction plan and the individual patient's needs.

[0061] In generating personalized correction models, by determining data mapping links, dependencies, and data embedding information, a mapping vector containing multiple vectors is generated. This enables the correction model to accurately reflect the data mapping relationships and component associations of correction needs, improving the accuracy and effectiveness of the correction model. This data-driven modeling approach can better adapt to the personalized needs of different patients, providing strong support for precision treatment.

[0062] Regarding the generation of resource allocation strategies, an optimization model incorporating multiple corrective constraints was constructed and solved using heuristic search algorithms, dynamic programming strategies, and constraint satisfaction algorithms. This achieved a rational allocation of data processing resources and minimized the migration cost of morphological matching chains. This scientifically sound resource allocation strategy improved resource utilization efficiency, shortened the time required to develop corrective plans, reduced treatment costs, and ensured the accuracy and reliability of the corrective process.

[0063] Furthermore, during the generation of the full spine contour model, multimodal data fusion, noise filtering, resolution enhancement, and inter-layer interpolation algorithms improved the quality of image data and the accuracy of the model, enabling the full spine contour model to more realistically reflect the actual morphology of the spine. Obtaining parameter distribution information, through constructing a parameter relationship graph and extracting weighted correlation paths, allows for in-depth analysis of the relationships between parameter nodes, providing more comprehensive information for model construction and optimization. The generation of morphological component sequences, through prioritization queues and dynamic sorting and combination, ensures the rationality and effectiveness of the component sequences, improving the efficiency and accuracy of model generation. Attached Figure Description

[0064] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 This is a flowchart of a method for establishing and applying a whole spinal morphological model based on medical images, provided by an embodiment of the present invention.

[0066] Figure 2 This is a schematic diagram illustrating the working principle of generating a morphological matching chain in response to personalized correction needs, provided in an embodiment of the present invention.

[0067] Figure 3 This is a schematic diagram illustrating the working principle of generating a target parameter set based on morphological difference information, as provided in an embodiment of the present invention.

[0068] Figure 4 This is a schematic diagram illustrating the working principle of generating a personalized correction model based on a morphological matching chain, as provided in an embodiment of the present invention.

[0069] Figure 5 This is a block diagram of a device for establishing and applying a full spinal morphological model based on medical images, provided in an embodiment of the present invention.

[0070] Figure 6 This is a schematic diagram of the structure of a device for establishing and applying a full spinal morphological model, provided in an embodiment of the present invention. Detailed Implementation

[0071] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0072] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0073] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0074] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0075] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0076] This invention provides a method for establishing and applying a whole-spine morphological model based on medical images. This method can be implemented using a whole-spine morphological model establishment and application device, which can be a terminal or a server. Figure 1 The flowchart shown illustrates the method for establishing and applying a whole-spine morphological model based on medical images. The processing flow of this method may include the following steps:

[0077] S1. Obtain the morphological features and medical imaging data of the patient's entire spine, construct a surface morphological parameter layer based on the morphological features, construct a medical imaging data layer based on the medical imaging data, and construct a three-dimensional morphological architecture of the entire spine based on the surface morphological parameter layer and the medical imaging data layer.

[0078] In one feasible implementation, morphological features and medical imaging data of the patient's entire spine are collected. The morphological features specifically include surface morphological parameters such as height, weight, spinal length, and curvature angle, constituting a surface morphological parameter layer; the medical imaging data uses a whole-spine MRI image sequence, constituting a medical imaging data layer.

[0079] Furthermore, the three-dimensional morphological architecture of the entire spine to be constructed includes a surface morphological parameter layer and a medical imaging data layer. The surface morphological parameter layer consists of height, weight, spinal length, and curvature angle, while the medical imaging data layer consists of a whole-spine MRI image sequence. The surface morphological parameter layer and the medical imaging data layer are matched through a data association algorithm. The medical imaging data layer is then reconstructed in three dimensions to generate a whole-spine contour model. The data association algorithm is used to match the surface morphological parameter layer and the medical imaging data layer to ensure the correspondence and correlation of the two data.

[0080] Optionally, S1 also includes:

[0081] The medical image data layer is reconstructed in three dimensions to generate a full spine contour model; the full spine contour model is used to generate a full spine morphological model.

[0082] The generation of the full spine contour model includes:

[0083] Multimodal data fusion was performed on the whole spine MRI image sequence in the medical imaging data layer to obtain fused three-dimensional image data.

[0084] The fused 3D image data is subjected to noise filtering and resolution enhancement to obtain optimized image data.

[0085] Based on the optimized image data, a full spine contour model is generated using an inter-slice interpolation algorithm.

[0086] In one feasible implementation, multimodal data fusion is first performed on the whole spine MRI image sequences in the medical imaging data layer, integrating image data from different sequences such as T1-weighted images, T2-weighted images, and fat-suppressed sequences to form three-dimensional image data containing multi-dimensional tissue information. Multimodal fusion is achieved through a registration algorithm, aligning the spatial coordinate systems of different image sequences to ensure the consistent position of the same anatomical structure in different sequences.

[0087] Furthermore, noise filtering and resolution enhancement are performed on the 3D image data. Noise filtering employs an adaptive median filtering algorithm, which dynamically adjusts the filtering window size based on the noise distribution in local areas. The calculation formula is as follows:

[0088] (1)

[0089] in, Represents the coordinates in the original image Pixel value at that location, For The dynamic filtering window centered on the center. These are the pixel values ​​output after filtering. This algorithm removes salt-and-pepper noise and Gaussian noise from the image, improving image clarity. Resolution enhancement processing employs a bicubic interpolation algorithm, which generates a high-resolution image by calculating the weighted sum of the 16 pixels surrounding the interpolation point, improving detail display capabilities and yielding optimized image data.

[0090] Furthermore, based on the optimized image data, a full spine contour model is generated using an inter-slice interpolation algorithm. The inter-slice interpolation algorithm calculates the pixel values ​​of intermediate slices based on the contour features of adjacent slice images. The specific steps are as follows: First, extract the spinal edge contour points from adjacent upper and lower slices, and determine corresponding point pairs using a contour matching algorithm; then, perform linear interpolation between each pair of corresponding points to generate contour points for intermediate slices; finally, connect the contour points from all slices to form a continuous three-dimensional surface, thus forming a full spine contour model. This model accurately reflects the three-dimensional geometric morphology of the spine, providing a basic geometric framework for establishing a full spine morphological model.

[0091] S2. Generate a full spine morphological model based on the three-dimensional morphological architecture of the entire spine.

[0092] In one feasible implementation, based on the constructed three-dimensional morphological architecture of the entire spine, further processing is performed to generate a morphological model of the entire spine. This model consists of two parts: a surface parameter model and an imaging data model. The surface parameter model is composed of multiple morphological index parameter nodes, used to characterize the patient's surface morphological features. The imaging data model is composed of multiple imaging data nodes and their spatial topological relationships, reflecting the anatomical structure and spatial distribution of the entire spine.

[0093] Optionally, step S2 above may include the following steps S21-S26:

[0094] S21. Based on the three-dimensional morphological architecture of the whole spine, determine the association status of each parameter node in the body surface morphology parameter layer and the spatial distribution status of each image data node in the medical image data layer.

[0095] In one feasible implementation, the logical relationships between various parameter nodes (such as height, weight, spinal length, curvature angle, etc.) need to be clearly defined in the body surface morphology parameter layer. For example, height and spinal length usually show a positive correlation, and the correlation coefficient between the two can be calculated using statistical methods to quantify the degree of correlation; curvature angle may have a potential correlation with weight, and the mapping relationship between the two needs to be established through clinical data.

[0096] For the medical imaging data layer, it is necessary to analyze the spatial distribution of each imaging data node (such as the imaging features of structures like vertebrae, intervertebral discs, and spinal cord) in the whole spinal MRI imaging sequence. For example, it is necessary to determine the coordinate position of each vertebra in three-dimensional space, the sequence connection relationship between adjacent vertebrae, and the specific location of the intervertebral disc in the intervertebral space.

[0097] S22. Obtain the parameter distribution information of the body surface morphology parameter layer and the image topology information of the medical image data layer based on the association status of each parameter node and the spatial distribution status of each image data node.

[0098] In one feasible implementation, the process of obtaining parameter distribution information includes:

[0099] S221. Construct a parameter relationship graph based on the association status of each parameter node in the body surface morphology parameter layer.

[0100] S222. Extract the weighted correlation paths between parameter nodes using a graph traversal algorithm.

[0101] S223. Calculate the distribution density and topological contribution of each parameter node based on the weighted associated path.

[0102] Specifically, in the process of acquiring parameter distribution information, a parameter relationship graph is first constructed. Each parameter node in the body surface morphology parameter layer is used as a graph node, and the correlation strength between nodes is used as the edge weight. For example, height and spine length nodes are used as vertices, and their correlation coefficient is used as the edge weight to construct the parameter relationship graph. A graph traversal algorithm (such as breadth-first search or depth-first search: breadth-first search uses a priority queue to implement Dijkstra's algorithm, expanding nodes by cumulative weight to ensure priority exploration of highly correlated paths; while depth-first search uses recursive depth limits (such as max_depth=10) and pruning strategies (terminating when the current path weight < threshold) to avoid invalid searches. This is suitable for mining long-range dependencies) is used to extract weighted correlation paths between parameter nodes, identifying key parameter nodes and their mutually influential paths (key path determination criteria: high-frequency paths, occurrence frequency > threshold (such as Top 5%); or high-weight paths: cumulative weight > ( The mean, (where the standard deviation is 1). Then, the distribution density and topological contribution of each parameter node are calculated based on the weighted association path: the distribution density reflects the tightness of the connection between parameter nodes in the graph, which can be quantified by the degree of the node (i.e., the number of edges connected to the node); the topological contribution measures the structural importance of the parameter node in the entire parameter network, which can be calculated using the PageRank algorithm or centrality analysis methods (such as betweenness centrality, proximity centrality) to determine the parameter nodes that have a greater impact on the overall morphological features.

[0103] To acquire image topology information for medical imaging data layers, it is necessary to analyze the spatial connectivity of each image data node based on the three-dimensional reconstruction results of MRI image sequences. For example, establishing the topological structure of the vertebral body sequence clarifies the sequential connectivity from the 1st cervical vertebra to the 5th lumbar vertebra; constructing the association between intervertebral discs and adjacent vertebral bodies to determine which two vertebral bodies each intervertebral disc is located between; and simultaneously, labeling the spatial positional relationship between the spinal cord and the vertebral bodies, such as the spinal cord being located within the spinal canal posterior to the vertebral bodies. Through the above analysis, an image topology network reflecting the anatomical structure of the entire spine is formed, providing a spatial structural basis for subsequent model construction.

[0104] S23. Obtain the attribute information of each parameter node and the attribute information of each image data node based on the parameter distribution information and image topology information.

[0105] In one feasible implementation, for parameter nodes in the body surface morphology parameter layer, attribute information includes parameter type (e.g., numerical, proportional), unit (e.g., centimeters, degrees), data source (e.g., clinical measurements, patient self-reports), and update frequency (e.g., static or dynamic parameters). For example, height and weight are numerical parameters, with units of centimeters and kilograms respectively, and are considered static parameters; bending angle may be a dynamic parameter, requiring measurement based on different body positions (e.g., standing, supine).

[0106] For image data nodes in a medical imaging data layer, attribute information includes imaging sequence type (such as T1WI, T2WI, FLAIR), slice thickness, pixel resolution, and grayscale range. For example, T1WI sequences are good at displaying anatomical structures, with a slice thickness of 3 mm and a pixel resolution of 512×512; T2WI sequences are sensitive to tissue edema, with a slice thickness of 2.5 mm and a pixel resolution of 384×384.

[0107] S24. Obtain the data resource information of each parameter node and the data resource information of each image data node.

[0108] In one feasible implementation, it is necessary to obtain the data resource information of each node, including data storage path (such as a specific folder in the hospital PACS system), access permissions (such as whether encrypted transmission is required), processing requirements (such as whether artifact removal processing is required), etc.

[0109] S25. Construct a body surface parameter model based on the attribute information and data resource information of each parameter node, and construct an image data model based on the attribute information and data resource information of each image data node.

[0110] In one feasible implementation, when constructing the body surface parameter model, the connection weights and dependencies between nodes are determined based on parameter nodes according to parameter distribution information, forming a network structure containing multiple morphological index parameter nodes. For example, strongly correlated parameter nodes such as height and spine length are connected through high-weight edges to highlight their role in synergistically representing spinal morphology.

[0111] When constructing an image data model, a three-dimensional structural model containing multiple image data nodes and their spatial topological relationships is established based on image topological information and the attribute information of image data nodes. For example, each vertebral node is arranged sequentially according to the vertebral body sequence, and the positional relationship between adjacent vertebrae is defined by spatial coordinates. At the same time, intervertebral disc nodes are embedded into the corresponding intervertebral spaces to form a complete spinal anatomical structure model.

[0112] S26. Construct a full spinal morphological model based on the body surface parameter model and the imaging data model.

[0113] In one feasible implementation, by integrating the body surface parameter model and the imaging data model, a full spinal morphological model that can comprehensively reflect the morphological characteristics and imaging data structure of the patient's entire spine is finally generated, providing basic data support for the analysis and processing of subsequent personalized correction needs.

[0114] S3. Obtain the patient's personalized correction needs and generate morphological difference information based on the personalized correction needs; generate a morphological matching chain based on the morphological difference information and the whole spine morphological model.

[0115] One feasible implementation method is, for example Figure 2 As shown, when a patient has personalized correction needs, the morphological differences corresponding to these needs are analyzed, and a morphological matching chain is generated by combining this with a full spinal morphological model. This matching chain is used to connect the various morphological nodes and components involved in the correction needs, providing path support for the generation of personalized correction models.

[0116] Optionally, generating the morphological matching chain based on morphological difference information and the whole spine morphological model in S3 may include the following steps S31-S33:

[0117] S31. Generate a target parameter set based on morphological difference information and the surface morphological parameter layer of the whole spine morphological model; wherein, the target parameter set includes multiple target morphological nodes, which are used to provide morphological matching resources required for correction needs.

[0118] Optionally, step S31 above may include the following steps S311-S316:

[0119] S311. Based on morphological difference information, multiple candidate morphological nodes are selected from the surface morphological parameter layer of the whole spine morphological model.

[0120] One feasible implementation method is, for example Figure 3 As shown, for patients' personalized correction needs (such as scoliosis correction, vertebral slippage reduction, etc.), candidate morphological nodes are screened from the surface morphological parameter layer based on morphological differences determined by clinical diagnosis (such as scoliosis angle exceeding the normal range, vertebral displacement distance, etc.). For example, if the correction need is angle correction for idiopathic scoliosis, morphological differences may include Cobb angle measurements, apical vertebra position, and vertebral segments involved in the scoliosis. Based on this, parameter nodes related to the scoliosis, such as the Cobb angle of the main curve, the Cobb angle of the compensatory curve, apical vertebral rotation, and spinal length differences, are extracted from the surface morphological parameter layer to form an initial set of candidate morphological nodes.

[0121] S312. Generate multiple candidate parameter sets based on multiple candidate morphological nodes.

[0122] In one feasible implementation, the generation method includes, but is not limited to: grouping by parameter correlation (e.g., grouping Cobb angle and apical vertebra position as one group, and spinal length difference and vertebral rotation as another group), grouping by correction stage (e.g., initial correction focuses on the Cobb angle of the main curve, mid-term focuses on compensatory curve balance, and late-term focuses on vertebral rotation improvement), and grouping by data type (e.g., numerical parameter node group, spatial location parameter node group). Each candidate parameter set contains different combinations of candidate morphological nodes. For example, candidate parameter set A includes the Cobb angle of the main curve and apical vertebra position, candidate parameter set B includes the Cobb angle of the main curve, apical vertebra rotation, and spinal length difference, and candidate parameter set C includes the Cobb angle of the compensatory curve and apical vertebra position, etc., to cover correction needs in different dimensions.

[0123] S313. Perform morphological evaluation and resource evaluation on each candidate parameter set to obtain the evaluation result of each candidate parameter set; wherein, the evaluation result includes scores of multiple evaluation dimensions; the multiple evaluation dimensions include parameter matching degree dimension, data integrity dimension, spatial consistency dimension and resource consumption dimension.

[0124] In one feasible implementation, a multi-dimensional evaluation is performed on each set of candidate parameters.

[0125] Specifically, the morphological assessment dimensions include:

[0126] ① Parameter matching degree: Determines the direct correlation between nodes in the candidate parameter set and the correction requirements. For example, for the correction requirement of scoliosis angle, the parameter matching degree of the main curve Cobb angle is significantly higher than that of the weight parameter.

[0127] ② Spatial Consistency: Analyze whether the spatial logic between candidate parameter nodes conforms to the anatomical structure of the spine. For example, the position of the apex vertebra and the curvature angle of the adjacent vertebrae should be spatially consistent. If there is a contradiction (such as the apex vertebra being labeled as T8, but the maximum curvature angle is T10), then the spatial consistency is low.

[0128] Furthermore, the dimensions of resource assessment include:

[0129] ① Data integrity: Check whether the data of each node in the candidate parameter set is complete and available. For example, if the vertebral rotation data contained in a candidate set is missing due to image quality issues, the data integrity score will be reduced.

[0130] ② Resource consumption: Evaluate the computational resources (such as memory and CPU time) and data storage costs required to process the set of candidate parameters. For example, parameter nodes that include three-dimensional spatial coordinates (such as cone position) typically consume more computational resources than single numerical parameters (such as Cobb angle).

[0131] Each dimension uses a standardized scoring system (e.g., 1-10 points), and is scored automatically or manually by the system according to preset rules. For example, parameter matching score = (number of directly related nodes in candidate nodes / total number of candidate nodes) × 10; data integrity score is calculated in reverse based on the proportion of missing data (e.g., 10 points for no missing data, 8 points for 20% missing data).

[0132] S314. Determine the weighting coefficient for each evaluation dimension based on individualized correction needs.

[0133] In one feasible implementation, the weight coefficients of each evaluation dimension are determined based on the clinical priority of the correction needs. For example, for urgent correction needs (such as severe scoliosis leading to respiratory dysfunction), the parameter matching weight can be set to 0.4, spatial consistency weight to 0.3, data integrity weight to 0.2, and resource consumption weight to 0.1; for non-urgent needs, the parameter matching weight can be appropriately reduced, and the resource consumption weight increased to optimize computational efficiency. The weight coefficients are set through clinical expert experience or preset rules to form a weight vector (e.g., [0.4, 0.3, 0.2, 0.1]).

[0134] S315. The evaluation results of each candidate parameter set are weighted based on the weight coefficients to obtain the comprehensive score of each candidate parameter set.

[0135] The evaluation results of each candidate parameter set are weighted based on weighting coefficients, and the formula is as follows:

[0136] (2)

[0137] For example, if candidate parameter set A has a parameter matching score of 8, spatial consistency score of 7, data integrity score of 9, and resource usage score of 6, with weighting coefficients of 0.4, 0.3, 0.2, and 0.1 respectively, then the comprehensive score is: 8×0.4+7×0.3+9×0.2+6×0.1=3.2+2.1+1.8+0.6=7.7.

[0138] S316. Based on the comprehensive score of each candidate parameter set, select the target parameter set from multiple candidate parameter sets.

[0139] In one feasible implementation, the optimal set of target parameters is selected by sorting the comprehensive scores from high to low. Assuming candidate sets A (7.7 points), B (6.5 points), and C (8.2 points), the candidate set C with the highest comprehensive score is selected as the target parameter set, which includes nodes such as the compensatory Cobb angle and the position of the apex vertebra, providing core morphological matching resources for subsequent correction schemes.

[0140] S32. Generate a morphological component sequence based on the target parameter set and personalized correction needs; wherein, the morphological component sequence includes multiple morphological components.

[0141] In one feasible implementation, the process of generating the morphological component sequence includes:

[0142] S321. Divide the priority queues according to the attribute information of each morphological node in the target parameter set.

[0143] S322. Dynamically sort and combine morphological components based on priority queues.

[0144] S323. Determine the order of morphological component sequences using an adaptive matching algorithm.

[0145] Specifically, after determining the target parameter set, a sequence of morphological components is generated based on this set and the correction requirements. First, a priority queue is formed according to the attribute information (such as parameter type and clinical priority) of each morphological node in the target parameter set. For example, the apical vertebra position, as a core parameter for spatial positioning, is given the highest priority; the compensatory Cobb angle, as a parameter reflecting overall balance, has the next highest priority. The morphological components (such as data acquisition components, parameter calculation components, and model prediction components) are dynamically sorted and combined through the priority queue: components corresponding to high-priority nodes (such as the apical vertebra position positioning component) are executed first to ensure the timeliness of processing key data.

[0146] Furthermore, an adaptive matching algorithm is used to determine the order of the morphological component sequences. The algorithm generates the optimal sequence based on the logical dependencies between components (e.g., the parameter calculation component must execute after the data acquisition component) and resource consumption (e.g., avoiding the concentrated execution of computationally intensive components), employing a greedy strategy or heuristic method. For example, the data acquisition component is executed first to obtain vertebral position image data, then the parameter calculation component is run to calculate the compensatory Cobb angle, and finally the model prediction component is called to generate a preliminary correction scheme.

[0147] S33. Generate a morphological matching chain based on the association between each target morphological node in the target parameter set and the morphological component sequence.

[0148] In one feasible implementation, a morphological matching chain is generated based on the relationships between morphological nodes in the target parameter set (such as the spatial relationship between the apical vertebra position and the compensatory Cobb angle) and the sequence of morphological components. The matching chain uses nodes and components as basic units, with directed edges connecting them to represent data flow and processing order. For example, the apical vertebra position node is connected to the parameter calculation component via the data acquisition component, and the compensatory Cobb angle data output by the parameter calculation component is passed to the model prediction component, forming a chain structure from node to component and back to node. This ensures that the morphological information involved in the correction needs can be processed and matched in a logical order, providing a clear path for the construction of a personalized correction model.

[0149] S4. Generate a personalized correction model for the patient based on the morphological matching chain.

[0150] One feasible implementation method is, for example Figure 4 As shown, a personalized correction model for the patient is further constructed based on the generated morphological matching chain. This model integrates relevant nodes and data from the morphological matching chain to form a customized correction plan tailored to the patient's specific correction needs.

[0151] Optionally, step S4 above may include the following steps S41-S47:

[0152] S41. Determine the source data node set and target data node set based on personalized correction needs.

[0153] In one feasible implementation, for the patient's personalized correction needs, a source data node set and a target data node set are clearly defined. The source data node set includes all relevant nodes reflecting the patient's current spinal morphology before correction, such as parameter nodes in the body surface morphology parameter layer (Cobb angle, spinal length, curvature angle, etc.) and imaging data nodes in the medical imaging data layer (vertebral body position, intervertebral disc signal intensity, etc.). The target data node set corresponds to the morphological indicators and imaging features expected to be achieved after correction, such as the corrected Cobb angle threshold, vertebral body alignment sequence, and intervertebral disc spatial position. Taking scoliosis correction as an example, the source data node set may include the main curve Cobb angle (current value 45°), apical vertebra position (T8), vertebral body rotation (Grade II), etc., while the target data node set is set as Cobb angle ≤ 15°, apical vertebra rotation ≤ Grade I, and restoration of spinal sequence symmetry, etc.

[0154] S42. Determine the data mapping link for personalized correction needs based on the source data node set and the target data node set.

[0155] In one feasible implementation, a data mapping link for the correction requirement is determined based on the source data node set and the target data node set. The data mapping link refers to the logical transformation path from the source data node to the target data node, requiring analysis of the morphological differences and spatial relationships between them. For example, for the requirement to correct the Cobb angle of the principal curve from 45° to 15°, the data mapping link needs to define the magnitude and direction of the angle adjustment (e.g., applying corrective force to the convex side) and the vertebral segments involved (e.g., T5-T12). For improving vertebral rotation, the axis of rotational repositioning (e.g., coronal plane, sagittal plane) and specific operational parameters (e.g., the rotational repositioning angle of each vertebra) need to be determined. By establishing a one-to-one or many-to-one mapping relationship between the source nodes and the target nodes, a complete data transformation logic is formed.

[0156] S43. Determine the dependencies between morphological components in the morphological matching chain based on the data mapping link.

[0157] In one feasible implementation, the dependencies between morphological components in the morphological matching chain are analyzed based on the data mapping link (determined according to the transformation logic required between specific source data and target data). Morphological components include data preprocessing components, parameter calculation components, model building components, and orthopedic scheme generation components.

[0158] Specifically, dependencies are divided into temporal dependencies and data dependencies: temporal dependencies refer to the order in which components are executed. For example, the data preprocessing component must first denoise the MRI images before the parameter calculation component can extract the vertebral body edge features. Data dependencies refer to the fact that subsequent components need the output data of preceding components. For example, the model building component needs to generate a 3D model based on the vertebral body position coordinates output by the parameter calculation component. By drawing a dependency graph, the execution order and data transmission path of each component are clarified, avoiding logical conflicts.

[0159] S44. Construct a correlation model between morphological components based on the dependencies between them.

[0160] In one feasible implementation, a correlation model is constructed based on dependencies between morphological components. The correlation model is represented as a directed acyclic graph (DAG), where nodes are morphological components and directed edges represent data flow and dependencies between components. For example, a chain-like correlation model can be formed by data preprocessing component → parameter calculation component → model building component → orthopedic scheme generation component, with the output of the previous component serving as the input of the next. For processing flows with branches (such as simultaneous surface parameter analysis and image data reconstruction), a parallel branch structure can be constructed, with data convergence achieved through synchronization nodes. The correlation model must ensure that the input and output formats of each component match; for example, the coordinate data output by the parameter calculation component must conform to the 3D modeling data format requirements of the model building component.

[0161] S45. Determine the data embedding information of the morphological matching chain based on the association model between each morphological component.

[0162] In one feasible implementation, data embedding information for morphological matching chains is determined based on an association model. This data embedding information includes format conversion rules, feature extraction methods, and status identifiers for the data during component processing. For example, in the data preprocessing component, DICOM format MRI images need to be converted to NIfTI format for subsequent processing, and the contour features of the vertebral body edges need to be extracted as input to the parameter calculation component. In the orthopedic scheme generation component, a status identifier (such as "to be adjusted" or "confirmed") needs to be added to each orthopedic force parameter for clinician review. The data embedding information also needs to define mechanisms for handling abnormal data, such as triggering data re-collection or skipping mechanisms when input data for a component is missing, ensuring the robustness of the model building process.

[0163] S46. Generate a mapping vector based on data embedding information; wherein the mapping vector includes a component embedding vector, a data mapping vector, and a correction receiving vector.

[0164] In one feasible implementation, the component embedding vector represents the position and functional attributes of each morphological component in the personalized correction model, and is represented by a multi-dimensional vector. For example, the vector dimension may include component type (data processing / model building / scheme generation), execution order number, resource consumption level (low / medium / high), etc. Taking the data preprocessing component as an example, its component embedding vector can be represented as [data processing, 1, low], indicating that this component is of the data processing type, the execution order is step 1, and the resource consumption level is low.

[0165] Data mapping vector: Describes the transformation relationship from source data nodes to target data nodes, including transformation functions and parameter sets. For example, for Cobb angle correction, the data mapping vector can define the transformation function as a linear transformation (target value = source value × correction coefficient), the parameter set as the correction coefficient (e.g., 0.33, to correct 45° to 15°), and constraints (correction coefficient range 0.2-0.5), etc.

[0166] Correction Receiving Vector (CRV): This represents the feedback mechanism for the correction effect, including the data type of feedback (e.g., postoperative Cobb angle measurement), feedback frequency (e.g., 1 week / 1 month / 3 months postoperatively), and evaluation indicators (e.g., correction maintenance rate, complication rate). The CRV is used for subsequent model validation and protocol adjustments, but only defines the structural framework during the model generation phase, without involving specific data.

[0167] S47. Generate a personalized correction model for the patient based on the mapping vector.

[0168] In one feasible implementation, the model is based on component embedding vectors, sequentially calling various morphological components. It uses data mapping vectors to drive the transformation of source data nodes to target data nodes, achieving a programmatic expression of the correction logic. A feedback interface is reserved using the correction receiving vector, providing expansion space for model iterative optimization. For example, the personalized correction model first processes the source data through data preprocessing and parameter calculation components to generate a 3D model containing the current spinal morphology. Then, based on the transformation functions and parameters in the data mapping vectors, the model building component simulates the changes in spinal morphology under orthopedic force, generating a predicted target 3D model. Finally, the correction plan generation component outputs a personalized correction plan containing orthopedic appliance parameters, force direction, and treatment plan based on the prediction results. This plan can receive clinical feedback data through the correction receiving vector interface for subsequent model updates and plan adjustments. The entire process ensures that the model closely matches the individual morphological characteristics of the patient, achieving precise customization of the correction plan.

[0169] S5. Generate a resource allocation strategy for patients based on a personalized correction model and a full spinal morphology model; the resource allocation strategy is used to allocate 3D modeling nodes and data processing resources for personalized correction needs to ensure the efficiency and accuracy of the correction process.

[0170] Optionally, step S5 above may include the following steps S51-S54:

[0171] S51. Extract morphological features and data requirements of personalized correction needs from the personalized correction model.

[0172] In one feasible implementation, morphological feature information and data requirement information of the correction needs are extracted from the personalized correction model. The morphological feature information covers the specific morphological parameters that need to be adjusted (such as Cobb angle, vertebral displacement distance, etc.), the involved spinal anatomical location (such as thoracic segment, lumbar segment, or specific vertebral segment), and the correction direction (such as scoliosis correction, lordosis angle adjustment, etc.); the data requirement information includes the type of medical imaging data required (such as T1WI, T2WI image sequences), resolution requirements, data update frequency (such as image acquisition requirements at different stages before, during, and after surgery), and data processing accuracy (such as voxel size for 3D modeling, number of decimal places retained in parameter calculation, etc.).

[0173] S52. Generate corrective constraints based on morphological feature information and data requirement information; wherein, the corrective constraints include source node constraints, target node constraints, component embedding constraints, processing time constraints, data consistency constraints, node resource constraints, and mapping bandwidth constraints of the morphological matching chain.

[0174] In one feasible implementation, correction constraints are generated based on the extracted morphological feature information and data requirement information. These constraints specifically include: source node constraints for the morphological matching chain, i.e., requirements for the data format (e.g., DICOM format MRI images), completeness (e.g., no missing layers in the image sequence), and timeliness (e.g., image data must be acquired within the last 3 months) of the source data nodes; target node constraints, i.e., regulations on the accuracy standards (e.g., Cobb angle measurement error ≤ 2° after correction), output format (e.g., 3D models must be in STL format), and clinically acceptable range (e.g., vertebral body reduction distance must reach more than 80% of anatomical reduction) of the target data nodes; component embedding constraints, i.e., requirements for the compatibility of the morphological components with the data interface of the full spinal morphological model (e.g., the parameter calculation component must be able to read pixel coordinate data from MRI images) and component operating environment matching (e.g., the 3D modeling component must run on a server supporting GPU acceleration); and processing time constraints, i.e., regulations on the processing time of each group. The time limits for data processing include: (e.g., the image reconstruction component must generate a full 3D model of the spine within 30 minutes); the total time limit for generating correction plans (e.g., from acquiring patient data to outputting a personalized plan, no more than 2 hours); data consistency constraints, ensuring data integrity during transmission (e.g., using encrypted transmission protocols to prevent data tampering); and consistent definitions across different component processing stages (e.g., maintaining uniformity in the measurement methods for curvature angles in surface parameter calculations and image analysis); node resource constraints, limiting the computational resources occupied by each parameter node or image data node (e.g., memory usage for a single vertebral 3D model ≤ 512MB); and storage resource quotas (e.g., the data storage period for a patient's full spine image is 5 years, with a single case data volume ≤ 2GB); and mapping bandwidth constraints, specifying the minimum bandwidth requirements for data transmission between different nodes or components (e.g., real-time image data transmission must be ≥ 100Mbps) to avoid processing delays due to insufficient bandwidth.

[0175] S53. Generate an optimization model based on the correction constraints; wherein the optimization model includes a first optimization model and a second optimization model; the first optimization model is used to minimize the allocation deviation of data processing resources; the second optimization model is used to minimize the migration cost of the morphological matching chain.

[0176] In one feasible implementation, an optimization model is generated based on the aforementioned correction constraints, including a first optimization model and a second optimization model. The first optimization model aims to minimize the allocation deviation of data processing resources. It establishes a resource allocation matrix, matching data processing resources (such as CPU core count, GPU computing power, and storage capacity) with various data processing tasks required for correction (such as image reconstruction, parameter calculation, and model simulation). A linear programming method is used to solve for the resource allocation scheme, minimizing the sum of the absolute values ​​of the deviations between the actual allocation of each resource and the theoretical values ​​of the task requirements. The second optimization model aims to minimize the migration cost of the morphological matching chain. It analyzes the deployment costs of each component in the morphological matching chain in different computing environments (such as local servers and cloud platforms), including hardware procurement costs, network transmission costs, and maintenance costs. By constructing a cost function and constraints, a dynamic programming method is used to determine the optimal deployment location of the components, minimizing the migration cost of the entire matching chain.

[0177] S54. Solve the optimization model to obtain the solution results; generate a resource allocation strategy for the patient based on the solution results and the full spine morphology model.

[0178] In one feasible implementation, the solution of the optimization model includes: traversing the candidate solution space of the morphological matching chain using a heuristic search algorithm, iteratively optimizing the candidate solutions based on a dynamic programming strategy, and selecting the optimal solution that meets the correction constraint conditions through a constraint satisfaction algorithm.

[0179] Specifically, when solving the optimization model, a heuristic search algorithm (such as genetic algorithm or particle swarm optimization algorithm) is first used to traverse the candidate solution space of the morphological matching chain, generating multiple potential resource allocation schemes. Heuristic search algorithms, by simulating biological evolution or swarm intelligence behavior, perform efficient searches in the solution space, avoiding getting trapped in local optima. Then, a dynamic programming strategy is used to iteratively optimize the candidate solutions, adjusting the search direction and parameters of the current stage based on the optimization results of the previous stage, gradually improving the quality of the solutions. For example, in genetic algorithms, a new generation of candidate solutions is generated through operations such as selection, crossover, and mutation, continuously optimizing the rationality of resource allocation. Finally, a constraint satisfaction algorithm (such as backtracking algorithm or constraint propagation algorithm) is used to screen the optimal solutions that meet the corrective constraints. Each candidate solution is checked to see if it satisfies all corrective constraints (such as resource allocation not exceeding the node resource limit, processing time meeting requirements, etc.), eliminating solutions that do not meet the conditions, and retaining the optimal solution that satisfies all constraints as the final resource allocation scheme.

[0180] Furthermore, based on the solution results of the optimization model and the full spine morphological model, a resource allocation strategy for patients is generated. This strategy clarifies the specific allocation scheme of 3D modeling nodes and data processing resources. For example, a server with GPU acceleration (IP address: 192.168.1.100) is designated as the full spine 3D modeling node, allocated 8 CPU cores, 16GB of memory, and 1TB of storage resources for processing patients' MRI image data; the parameter calculation task is decomposed into multiple sub-tasks and distributed to 3 computing nodes (node ​​A, node B, and node C) in the cloud cluster for parallel processing, with each node allocated 4 CPU cores and 8GB of memory; the image data transmission path is set as hospital PACS system → local preprocessing node → cloud computing node, with a guaranteed transmission bandwidth of 200Mbps; resource usage priorities are defined, and when multiple patients request resources simultaneously, scheduling is based on the urgency of the correction needs (e.g., priority for emergency surgery patients) and resource consumption (prioritizing tasks with low resource consumption). The resource allocation strategy should also include a dynamic adjustment mechanism. When the resource utilization of a node is detected to exceed the threshold (such as CPU utilization > 80%), resource expansion or task migration should be automatically triggered to ensure the efficient operation of the personalized correction model and the smooth realization of correction requirements.

[0181] In this embodiment of the invention, based on the patient's full spinal morphological characteristics and medical imaging data, a three-dimensional morphological architecture is constructed, including a layer of surface morphological parameters (height, weight, spinal length, curvature angle) and a layer of medical imaging data (full spinal MRI image sequence). A full spinal contour model is generated through three-dimensional reconstruction; subsequently, a full spinal morphological model composed of the surface parameter model and the imaging data model is generated. In response to personalized correction needs, a personalized correction model is generated based on morphological difference information and a morphological matching chain generated by the model, combined with a model generation resource allocation strategy to optimize resource allocation. The method also includes key steps such as multimodal data fusion, parameter evaluation and screening, and optimized model solving, achieving accurate full spinal modeling and personalized correction, improving the scientific rigor and resource utilization efficiency of spinal disease diagnosis and treatment.

[0182] Figure 5 This is a block diagram illustrating a device for establishing and applying a full-spine morphological model based on medical images, according to an exemplary embodiment. The device is used for a method of establishing and applying a full-spine morphological model based on medical images. (Refer to...) Figure 5 The device includes a full-spine three-dimensional morphological architecture construction module 310, a full-spine morphological model generation module 320, a morphological matching chain generation module 330, a personalized correction model generation module 340, and an output module 350. Among them:

[0183] The whole spine three-dimensional morphological architecture construction module 310 is used to acquire the morphological features and medical imaging data of the patient's whole spine, construct a body surface morphological parameter layer based on the morphological features, construct a medical imaging data layer based on the medical imaging data, and construct a whole spine three-dimensional morphological architecture based on the body surface morphological parameter layer and the medical imaging data layer.

[0184] The full spine morphology model generation module 320 is used to generate a full spine morphology model based on the three-dimensional morphological architecture of the full spine.

[0185] The morphological matching chain generation module 330 is used to obtain the patient's personalized correction needs and generate morphological difference information based on the personalized correction needs; and to generate a morphological matching chain based on the morphological difference information and the whole spinal morphological model.

[0186] Personalized correction model generation module 340 is used to generate a personalized correction model for the patient based on the morphological matching chain.

[0187] Output module 350 is used to generate a resource allocation strategy for patients based on a personalized correction model and a full spinal morphology model; wherein, the resource allocation strategy is used to allocate 3D modeling nodes and data processing resources for personalized correction needs.

[0188] In this embodiment of the invention, based on the patient's full spinal morphological characteristics and medical imaging data, a three-dimensional morphological architecture is constructed, including a layer of surface morphological parameters (height, weight, spinal length, curvature angle) and a layer of medical imaging data (full spinal MRI image sequence). A full spinal contour model is generated through three-dimensional reconstruction; subsequently, a full spinal morphological model composed of the surface parameter model and the imaging data model is generated. In response to personalized correction needs, a personalized correction model is generated based on morphological difference information and a morphological matching chain generated by the model, combined with a model generation resource allocation strategy to optimize resource allocation. The method also includes key steps such as multimodal data fusion, parameter evaluation and screening, and optimized model solving, achieving accurate full spinal modeling and personalized correction, improving the scientific rigor and resource utilization efficiency of spinal disease diagnosis and treatment.

[0189] Figure 6 This is a schematic diagram of the structure of a device for establishing and applying a full spinal morphological model, as provided in an embodiment of the present invention. Figure 6 As shown, the equipment for establishing and applying a full spine morphological model may include the above-mentioned... Figure 5 The illustrated device is a whole-spine morphological model creation and application device based on medical images. Optionally, the whole-spine morphological model creation and application device 410 may include a first processor 2001.

[0190] Optionally, the whole spine morphological model establishment and application device 410 may also include a memory 2002 and a transceiver 2003.

[0191] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0192] The following is combined Figure 6 A detailed introduction to each component of the whole spine morphological model establishment and application device 410 is provided below:

[0193] The first processor 2001 is the control center of the whole spine morphological model establishment and application device 410. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0194] Optionally, the first processor 2001 can perform various functions of the full spine morphological model establishment and application device 410 by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0195] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 6 CPU0 and CPU1 are shown in the diagram.

[0196] In a specific implementation, as one example, the whole spine morphological model establishment and application device 410 may also include multiple processors, for example... Figure 6 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0197] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0198] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently, and its interface circuit with the application device 410 may be established through a full-spine morphological model. Figure 6 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0199] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0200] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 6 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0201] Optionally, the transceiver 2003 can be integrated with the first processor 2001 or exist independently, and an interface circuit with the application device 410 can be established through a full spine morphological model. Figure 6 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0202] It should be noted that, Figure 6 The structure of the full spine morphology model building and application device 410 shown in the figure does not constitute a limitation on the router. The actual knowledge structure recognition device may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0203] Furthermore, the technical effects of the whole spine morphological model establishment and application device 410 can be referred to the technical effects of the whole spine morphological model establishment and application method based on medical imaging described in the above method embodiments, and will not be repeated here.

[0204] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0205] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0206] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0207] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0208] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0209] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0210] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0211] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0212] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0213] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0214] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for establishing and applying a whole-spine morphological model based on medical imaging, characterized in that, The method includes: S1. Obtain the morphological features and medical imaging data of the patient's entire spine, construct a body surface morphological parameter layer based on the morphological features, construct a medical imaging data layer based on the medical imaging data, and construct a three-dimensional morphological architecture of the entire spine based on the body surface morphological parameter layer and the medical imaging data layer. S2. Generate a full spine morphological model based on the aforementioned three-dimensional morphological architecture of the entire spine; S3. Obtain the patient's personalized correction needs and generate morphological difference information based on the personalized correction needs; generate a morphological matching chain based on the morphological difference information and the whole spine morphological model; S4. Generate a personalized correction model for the patient based on the morphological matching chain; S5. Generate a resource allocation strategy for the patient based on the personalized correction model and the whole spinal morphology model; wherein, the resource allocation strategy is used to allocate three-dimensional modeling nodes and data processing resources for the personalized correction needs.

2. The method for establishing and applying a whole-spine morphological model based on medical imaging according to claim 1, characterized in that, The generation of a full-spine morphological model based on the three-dimensional morphological architecture of the entire spine in S2 includes: S21. Based on the three-dimensional morphological architecture of the whole spine, determine the association status of each parameter node in the body surface morphology parameter layer and the spatial distribution status of each image data node in the medical image data layer. S22. Obtain the parameter distribution information of the body surface morphology parameter layer and the image topology information of the medical image data layer based on the association status of each parameter node and the spatial distribution status of each image data node. S23. Obtain the attribute information of each parameter node and the attribute information of each image data node based on the parameter distribution information and the image topology information. S24. Obtain the data resource information of each parameter node and the data resource information of each image data node; S25. Construct a body surface parameter model based on the attribute information and data resource information of each parameter node, and construct an image data model based on the attribute information and data resource information of each image data node. S26. Construct a full spinal morphological model based on the surface parameter model and the imaging data model.

3. The method for establishing and applying a whole-spine morphological model based on medical imaging according to claim 1, characterized in that, The generation of a morphological matching chain based on the morphological difference information and the whole spine morphological model in S3 includes: S31. Generate a target parameter set based on the morphological difference information and the surface morphological parameter layer of the whole spine morphological model; wherein, the target parameter set includes multiple target morphological nodes; S32. Generate a morphological component sequence based on the target parameter set and personalized correction needs; wherein, the morphological component sequence includes multiple morphological components; S33. Generate a morphological matching chain based on the association between each target morphological node in the target parameter set and the morphological component sequence.

4. The method for establishing and applying a whole-spine morphological model based on medical imaging according to claim 3, characterized in that, The generation of a target parameter set in S31 based on the morphological difference information and the surface morphological parameter layer of the whole spine morphological model includes: S311. Based on the morphological difference information, select multiple candidate morphological nodes from the body surface morphological parameter layer of the whole spine morphological model. S312. Generate multiple candidate parameter sets based on the multiple candidate morphological nodes; S313. Perform morphological evaluation and resource evaluation on each candidate parameter set to obtain the evaluation result of each candidate parameter set; wherein, the evaluation result includes scores of multiple evaluation dimensions; the multiple evaluation dimensions include parameter matching degree dimension, data integrity dimension, spatial consistency dimension and resource consumption dimension; S314. Determine the weighting coefficient for each evaluation dimension based on the personalized correction needs. S315. The evaluation results of each candidate parameter set are weighted based on the weight coefficients to obtain a comprehensive score for each candidate parameter set. S316. Based on the comprehensive score of each candidate parameter set, select the target parameter set from the multiple candidate parameter sets.

5. The method for establishing and applying a whole-spine morphological model based on medical imaging according to claim 1, characterized in that, The process of generating a personalized correction model for the patient based on the morphological matching chain in S4 includes: S41. Determine the source data node set and the target data node set based on the personalized correction requirements; S42. Determine the data mapping link for personalized correction needs based on the source data node set and the target data node set; S43. Determine the dependencies between each morphological component in the morphological matching chain based on the data mapping link; S44. Construct an association model between the morphological components based on the dependencies between them; S45. Determine the data embedding information of the morphological matching chain based on the association model between the morphological components; S46. Generate a mapping vector based on the data embedding information; wherein the mapping vector includes a component embedding vector, a data mapping vector, and a correction receiving vector; S47. Generate a personalized correction model for the patient based on the mapping vector.

6. The method for establishing and applying a whole-spine morphological model based on medical imaging according to claim 1, characterized in that, The resource allocation strategy for generating patients based on the personalized correction model and the whole spine morphology model in S5 includes: S51. Extract morphological feature information and data requirement information of personalized correction needs from the personalized correction model; S52. Generate correction constraints based on the morphological feature information and data requirement information; wherein, the correction constraints include source node constraints, target node constraints, component embedding constraints, processing time constraints, data consistency constraints, node resource constraints, and mapping bandwidth constraints of the morphological matching chain. S53. Generate an optimization model based on the correction constraints; wherein the optimization model includes a first optimization model and a second optimization model; the first optimization model is used to minimize the allocation deviation of data processing resources; the second optimization model is used to minimize the migration cost of the morphological matching chain; S54. Solve the optimization model to obtain the solution results; generate a resource allocation strategy for the patient based on the solution results and the whole spine morphology model.

7. The method for establishing and applying a whole-spine morphological model based on medical imaging according to claim 1, characterized in that, The S1 further includes: The medical image data layer is reconstructed in three dimensions to generate a full spine contour model; wherein, the full spine contour model is used to generate a full spine morphological model; The generation of the full spine contour model includes: Multimodal data fusion was performed based on the whole spine MRI image sequence in the medical imaging data layer to obtain fused three-dimensional image data; The fused 3D image data is subjected to noise filtering and resolution enhancement processing to obtain optimized image data. Based on the optimized image data, a full spine contour model is generated using an inter-layer interpolation algorithm.

8. A device for establishing and applying a whole-spine morphological model based on medical images, wherein the device is used to implement the method for establishing and applying a whole-spine morphological model based on medical images as described in any one of claims 1-7, characterized in that, The device includes: A three-dimensional morphological architecture construction module for the whole spine is used to acquire the morphological features and medical imaging data of the patient's whole spine, construct a body surface morphological parameter layer based on the morphological features, construct a medical imaging data layer based on the medical imaging data, and construct a three-dimensional morphological architecture for the whole spine based on the body surface morphological parameter layer and the medical imaging data layer. A full spine morphological model generation module is used to generate a full spine morphological model based on the three-dimensional morphological architecture of the full spine. The morphological matching chain generation module is used to obtain the patient's personalized correction needs, generate morphological difference information based on the personalized correction needs, and generate a morphological matching chain based on the morphological difference information and the whole spine morphological model. A personalized correction model generation module is used to generate a personalized correction model for the patient based on the morphological matching chain. The output module is used to generate a resource allocation strategy for the patient based on the personalized correction model and the whole spine morphology model; wherein the resource allocation strategy is used to allocate three-dimensional modeling nodes and data processing resources for the personalized correction needs.

9. A device for establishing and applying a full spine morphological model, characterized in that, The equipment for establishing and applying the whole spine morphological model includes: processor; A memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be invoked by a processor to execute the method as described in any one of claims 1 to 7.

Images