CN-121983233-A - Digital twinning-based knee joint personalized rehabilitation scheme generation method and system

CN121983233ACN 121983233 ACN121983233 ACN 121983233ACN-121983233-A

Abstract

The invention provides a method and a system for generating a personalized knee joint rehabilitation scheme based on digital twinning, which are used for acquiring multidimensional characteristic data of a patient, constructing a patient rehabilitation data set, constructing a knee joint injury knowledge graph based on the patient rehabilitation data set, constructing a knee joint tiny injury model based on the knee joint injury knowledge graph, accurately identifying functional weak links and injury degrees of the patient during rehabilitation training, providing a data basis for the subsequent establishment of a personalized scheme aiming at the illness state of the patient, carrying out emotion extraction on voice feedback data in the patient rehabilitation data set, acquiring a patient emotion index, carrying out rehabilitation efficiency diagnosis according to the patient emotion index, generating a knee joint initial rehabilitation scheme, carrying out efficacy evaluation and scheme optimization on the knee joint initial rehabilitation scheme, generating a knee joint rehabilitation scheme, and dynamically adjusting the rehabilitation scheme according to the personalized requirement of the patient while meeting the personalized requirement of the patient, thereby improving the accuracy and efficiency of the generated rehabilitation scheme.

Inventors

XUE JING
DU JUNJIE
CHU DONG
LI XIAOJIE
GAN LU
PENG YE

Assignees

中国人民解放军空军特色医学中心

Dates

Publication Date: 20260505
Application Date: 20251208

Claims (10)

1. The method for generating the personalized rehabilitation scheme of the knee joint based on digital twinning is characterized by comprising the following steps of: Acquiring multidimensional characteristic data of a patient, and constructing a patient rehabilitation data set; constructing a knee joint injury knowledge graph based on the patient rehabilitation data set; Constructing a knee joint micro injury model based on the knee joint injury knowledge graph; extracting emotion from voice feedback data in a patient rehabilitation data set to obtain a patient emotion index; Diagnosing rehabilitation efficiency according to the emotion index of the patient, and generating an initial rehabilitation scheme of the knee joint; And performing efficacy evaluation and scheme optimization on the initial knee joint rehabilitation scheme to generate a knee joint rehabilitation scheme.
2. The method for generating a personalized rehabilitation scheme for knee joints based on digital twinning according to claim 1, wherein constructing a miniature knee joint injury model based on the knowledge graph of knee joint injury comprises: inputting data corresponding to the knee joint biomechanical nodes and the knee joint image nodes contained in the knee joint injury knowledge graph into a deep learning neural network to generate knee joint injury simulation data; Constructing a knee joint injury preliminary mapping rule according to time sequence association relations and causal association relations among all nodes contained in a knee joint injury node set in a knee joint injury knowledge graph, and carrying out iterative correction on the knee joint injury preliminary mapping rule by taking rehabilitation label node corresponding data contained in the knee joint injury knowledge graph as a reference until the similarity between a mapping result corresponding to the knee joint injury preliminary mapping rule and the rehabilitation label node corresponding data is greater than a preset similarity threshold value, so as to generate the knee joint injury mapping rule; And constructing an original knee joint micro-damage model based on the knee joint damage simulation data and the knee joint damage mapping rule, performing simulation verification by adopting a finite element algorithm, and performing optimization correction on the original knee joint micro-damage model according to a simulation verification result to generate the knee joint micro-damage model.
3. The digital twinning-based knee joint personalized rehabilitation scheme generation method according to claim 1, wherein emotion extraction is performed on voice feedback data in a patient rehabilitation data set to obtain a patient emotion index, comprising: Acquiring text feedback data corresponding to a voice feedback node contained in a knee joint injury knowledge graph and corresponding voice audio original data; Extracting keywords from text feedback data based on a BERT model, generating a patient emotion keyword set, outputting emotion category probability distribution corresponding to the emotion keywords according to positive emotion and negative emotion categories of each emotion keyword in the patient emotion keyword set based on a Softmax activation function, and outputting a probability value of the positive emotion as emotion polarity score; Extracting features of voice audio original data to obtain an audio feature set, and combining corresponding text feedback data to obtain emotion tendency data of a patient; Obtaining corresponding data of knee joint biomechanical nodes in a knee joint injury knowledge graph, comparing the corresponding data with a preset fatigue threshold and a load state threshold, and generating physiological load data of a patient according to a comparison result; and constructing a multi-mode fusion network based on an attention mechanism, and carrying out self-adaptive weighting processing on the emotion polarity fraction, the patient emotion tendency data and the patient physiological load data through a dynamic weight distribution strategy to generate a patient emotion index.
4. The method for generating a personalized rehabilitation regimen for knee joints based on digital twinning according to claim 1, wherein the diagnosis of rehabilitation efficiency is performed according to the emotional index of the patient, and the generation of an initial rehabilitation regimen for knee joints comprises: based on the patient rehabilitation data set, matching a historical optimal rehabilitation scheme in a preset optimal rehabilitation scheme library, sending the historical optimal rehabilitation scheme to a user side, monitoring the emotion index of the patient when the patient executes the historical optimal rehabilitation scheme in real time, and synchronously updating the patient rehabilitation data set; If the emotion index of the patient is within the preset emotion index threshold interval, continuously monitoring the emotion index of the patient, and synchronously updating a rehabilitation data set of the patient; If the emotion index of the patient is not located in the preset emotion index threshold value interval, inputting the updated patient rehabilitation data set into a knee joint micro-damage model, converting the patient rehabilitation data set into knee joint damage simulation data corresponding to the knee joint micro-damage model, and positioning a rehabilitation training weak area of the patient according to knee joint damage mapping rules contained in the knee joint micro-damage model to generate rehabilitation training optimization data; the rehabilitation training optimization data at least comprise the position of a rehabilitation training weak area, the stress tolerance upper limit and the current damage degree; Analyzing the rehabilitation training optimizing data according to a near-end strategy optimizing algorithm to obtain a rehabilitation training scheme adjusting decision; and generating an initial knee joint rehabilitation scheme based on the rehabilitation training optimizing data and the rehabilitation training scheme adjustment decision.
5. The method for generating a personalized rehabilitation solution for knee joints based on digital twinning according to claim 4, wherein the analyzing the rehabilitation training optimizing data according to the proximal strategy optimizing algorithm to obtain the rehabilitation training solution adjustment decision comprises: Based on rehabilitation training optimization data, constructing a state space by combining rehabilitation label nodes in the knee joint injury knowledge graph and corresponding data of knee joint biomechanical nodes; And setting a reward function by taking the stress load of the weak area of the rehabilitation training as an optimization target, and performing iterative training through a near-end strategy optimization algorithm to generate a rehabilitation training scheme adjustment decision.
6. The method for generating a digital twinning-based knee joint personalized rehabilitation scheme according to claim 1, wherein performing performance evaluation and scheme optimization on a knee joint initial rehabilitation scheme to generate the knee joint rehabilitation scheme comprises: Inputting a knee joint initial rehabilitation scheme into a knee joint micro damage model, and carrying out simulation prediction by adopting a model prediction control algorithm to obtain a simulation prediction result, wherein the simulation prediction result at least comprises predicted rehabilitation effect data and predicted safety risk data, and the predicted rehabilitation effect data at least comprises stress reduction amplitude of a damaged area, knee joint moving range improvement rate and knee joint moving range improvement curve; verifying the simulation prediction result by combining causal association relations among all nodes in a knee joint injury node set contained in the knee joint injury knowledge graph; If the simulation prediction result does not pass the verification, carrying out parameter adjustment on the initial rehabilitation scheme of the knee joint until the simulation prediction result passes the verification; and if the simulation prediction result passes the verification, the corresponding knee joint initial rehabilitation scheme is used as the knee joint rehabilitation scheme to be sent to the user side.
7. The method for generating a personalized rehabilitation scheme for knee joint based on digital twinning according to claim 6, wherein verifying the simulation prediction result by combining causal association relations among nodes in a knee joint injury node set contained in a knee joint injury knowledge graph comprises: Acquiring causal association relation between knee joint biomechanical nodes and rehabilitation label nodes in a knee joint injury knowledge graph, and defining the causal association relation as a rehabilitation association rule; taking the stress reduction amplitude of the damaged area as a verification index, and combining with a rehabilitation association rule to verify the damage repair efficiency; taking the knee joint range of motion improvement rate as a check index, and combining with a rehabilitation association rule to check the stability of the rehabilitation effect; taking the knee joint movement range improvement curve as a verification index, and carrying out rehabilitation effect persistence verification by combining with a rehabilitation association rule; and acquiring causal association relation between knee joint biomechanical nodes and voice feedback nodes in the knee joint injury knowledge graph, and performing rehabilitation training tolerance verification.
8. The method for generating a digital twinning-based knee joint personalized rehabilitation scheme according to claim 1, wherein acquiring multi-dimensional characteristic data of a patient and constructing a patient rehabilitation data set comprises: collecting knee joint state data of a patient during rehabilitation exercise, and generating a continuous knee joint state sequence; Extracting characteristics of the knee joint state sequence to obtain knee joint biomechanical data and knee joint movement data; The method comprises the steps of synchronously collecting corresponding voice information of a patient during rehabilitation exercise, and generating voice feedback data of the patient, wherein the voice feedback data at least comprise subjective feeling data, pain description data and compliance feedback data of the patient; Extracting characteristics of a knee joint pathology report of a patient to obtain knee joint image data and knee joint diagnosis data, wherein the knee joint image data at least comprises knee joint structure data, and the knee joint diagnosis data at least comprises damage degree data and damage rehabilitation stage data; And constructing a patient rehabilitation data set based on the knee biomechanical data and the knee movement data, the voice feedback data, the knee image data and the knee diagnosis data.
9. The method for generating a digital twinning-based knee joint personalized rehabilitation scheme according to claim 1, wherein constructing a knee joint injury knowledge graph based on the patient rehabilitation data set comprises: Defining the biomechanical data and the movement data of the knee joint in the patient rehabilitation data set as knee joint biomechanical nodes, defining the voice feedback data in the patient rehabilitation data set as voice feedback nodes, defining the knee joint image data in the patient rehabilitation data set as knee joint image nodes, defining the knee joint diagnosis data in the patient rehabilitation data set as rehabilitation label nodes, and generating a knee joint injury node set; Performing time stamp alignment on the data corresponding to each node in the knee joint injury node set to obtain a time sequence association relationship among each node in the knee joint injury node set; Acquiring causal association relations among all nodes in the knee joint injury node set by adjusting data corresponding to all nodes in the knee joint injury node set and monitoring data changes of nodes which are not subjected to data adjustment in the knee joint injury node set; And constructing a knee joint injury knowledge graph based on the time sequence association relationship and the causal association relationship among all nodes in the knee joint injury node set.
10. The utility model provides a knee joint personalized rehabilitation scheme generation system based on digital twin which is characterized in that includes: The data acquisition module is used for acquiring multidimensional characteristic data of the knee joint injury patient and constructing a patient rehabilitation data set; the map construction module is used for constructing a knee joint injury knowledge map according to the patient rehabilitation data set; The model construction module is used for constructing a knee joint micro-injury model according to the knee joint injury knowledge graph; The emotion extraction module is used for extracting emotion from the voice feedback data in the patient rehabilitation data set to obtain a patient emotion index; the scheme generation module is used for diagnosing rehabilitation efficiency according to the emotion indexes of the patient and generating an initial knee joint rehabilitation scheme; The scheme optimization module is used for performing efficiency evaluation and scheme optimization on the initial knee joint rehabilitation scheme to generate a knee joint rehabilitation scheme.

Description

Digital twinning-based knee joint personalized rehabilitation scheme generation method and system Technical Field The invention relates to the technical field of data processing, in particular to a method and a system for generating a personalized rehabilitation scheme of a knee joint based on digital twinning. Background The formulation of knee joint rehabilitation schemes is an important research direction in the field of medical health, is directly related to the improvement of life quality of patients and the recovery of exercise functions, and along with the aging of population and the increase of exercise injury, the demands of personalized rehabilitation schemes are increasingly urgent. The effective rehabilitation scheme can help the patient to quickly recover the joint movement capability, and reduce the risk of secondary injury. However, the existing knee joint rehabilitation method still has a certain limitation in meeting the individual requirements of patients and the dynamic adjustment of rehabilitation schemes. On one hand, the existing knee joint rehabilitation method depends on a standardized training plan, the training plan is usually formulated based on experience of doctors or a general template, accurate capture of real-time states of patients is lacking, individual differences and dynamic changes of the patients are difficult to fully consider, the method cannot solve specific weak links in knee joint functions of the patients in a targeted manner when facing complex rehabilitation demands, on the other hand, the acceptance degree and execution willingness difference of the patients on training projects are large, and the existing method often lacks an effective mechanism to integrate subjective feedback of the patients while evaluating a rehabilitation scheme, so that the execution effect of the rehabilitation scheme is poor. Disclosure of Invention In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a method for generating a personalized rehabilitation regimen for a knee joint based on digital twinning, the method comprising: Acquiring multidimensional characteristic data of a patient, and constructing a patient rehabilitation data set; constructing a knee joint injury knowledge graph based on the patient rehabilitation data set; Constructing a knee joint micro injury model based on the knee joint injury knowledge graph; extracting emotion from voice feedback data in a patient rehabilitation data set to obtain a patient emotion index; Diagnosing rehabilitation efficiency according to the emotion index of the patient, and generating an initial rehabilitation scheme of the knee joint; And performing efficacy evaluation and scheme optimization on the initial knee joint rehabilitation scheme to generate a knee joint rehabilitation scheme. As a further scheme of the invention, constructing the knee joint micro-injury model based on the knee joint injury knowledge graph comprises the following steps: inputting data corresponding to the knee joint biomechanical nodes and the knee joint image nodes contained in the knee joint injury knowledge graph into a deep learning neural network to generate knee joint injury simulation data; Constructing a knee joint injury preliminary mapping rule according to time sequence association relations and causal association relations among all nodes contained in a knee joint injury node set in a knee joint injury knowledge graph, and carrying out iterative correction on the knee joint injury preliminary mapping rule by taking rehabilitation label node corresponding data contained in the knee joint injury knowledge graph as a reference until the similarity between a mapping result corresponding to the knee joint injury preliminary mapping rule and the rehabilitation label node corresponding data is greater than a preset similarity threshold value, so as to generate the knee joint injury mapping rule; And constructing an original knee joint micro-damage model based on the knee joint damage simulation data and the knee joint damage mapping rule, performing simulation verification by adopting a finite element algorithm, and performing optimization correction on the original knee joint micro-damage model according to a simulation verification result to generate the knee joint micro-damage model. As a further aspect of the present invention, emotion extraction is performed on voice feedback data in a patient rehabilitation data set to obtain a patient emotion index, including: Acquiring text feedback data corresponding to a voice feedback node contained in a knee joint injury knowledge graph and corresponding voice audio original data; Extracting keywords from text feedback data based on a BERT model, generating a patient emotion keyword set, outputting emotion category probability distribution corresponding to the emotion keywords according to positive emotion and