Search

CN-122025007-A - Muscle injury rehabilitation treatment method and system based on multi-mode data fusion

CN122025007ACN 122025007 ACN122025007 ACN 122025007ACN-122025007-A

Abstract

The application relates to a muscle injury rehabilitation therapy method and a system based on multi-mode data fusion, wherein the method comprises the following steps of S1, obtaining data comprising myoelectricity, electroencephalogram and gesture data, S2, carrying out fusion evaluation based on the multi-mode data to obtain multi-mode evaluation scores, S3, generating and executing an individual rehabilitation therapy scheme based on the multi-mode evaluation scores, S4, triggering and re-executing the steps of S1 and S2 when the rehabilitation state is detected to reach a preset threshold value, and updating the individual rehabilitation therapy scheme based on the newly generated multi-mode evaluation scores. According to the application, accurate assessment of muscle contraction state and injury degree is realized through multidimensional data fusion, a dynamic treatment scheme is generated by combining a personalized adaptation mechanism, and the rehabilitation progress is tracked and optimized through full-period data. The method solves the problems of single evaluation dimension, insufficient personalized adaptation and lack of dynamic adjustment of the treatment scheme in the prior art, and is suitable for rehabilitation scenes of various tissue injuries such as bones, ligaments, cartilages, skeletal muscles and the like.

Inventors

  • Zheng Jialu
  • MA JUN
  • Luo Haoyun
  • LI YUTING

Assignees

  • 重庆医科大学

Dates

Publication Date
20260512
Application Date
20260130

Claims (10)

  1. 1. The muscle injury rehabilitation treatment method based on multi-mode data fusion is characterized by comprising the following steps of: S1, acquiring myoelectric signals, electroencephalogram signals and posture data of a target object, and preprocessing to obtain myoelectric characteristic data, electroencephalogram index data and posture parameter data, so as to form multi-mode data; s2, fusion evaluation is carried out based on the multi-mode data so as to obtain multi-mode evaluation scores; S3, generating and executing an individual rehabilitation therapy scheme based on the multi-mode evaluation score and the personalized adaptation parameters of the target object; And S4, recording rehabilitation progress data of the target object, judging a rehabilitation state based on the rehabilitation progress data, triggering and re-executing the steps S1 and S2 when the rehabilitation state is detected to reach a preset threshold value, and updating an individual rehabilitation treatment scheme based on the newly generated multi-mode evaluation score.
  2. 2. The method according to claim 1, wherein the step S2 comprises: s21, extracting respective time stamps in the multi-modal data, and mapping the multi-modal data sampled in unequal periods to a preset reference time sequence axis through an interpolation algorithm so that each reference time point can correspond to a group of multi-modal data; s22, respectively distributing a sampling period adaptation coefficient for the multi-mode data mapped to the reference time sequence axis; S23, constructing a three-dimensional mode evaluation model comprising an myoelectric mode, an electroencephalogram mode and a gesture mode, respectively distributing mode weights for each mode, respectively distributing sub-weights for characteristic indexes under the corresponding modes, wherein the sum of the sub-weights of all the characteristic indexes under each mode is equal to the mode weight of the corresponding mode; S24, respectively calculating and obtaining single-mode evaluation values of corresponding modes based on corresponding data of the characteristic indexes under each mode and sub-weights of the corresponding characteristic indexes, and multiplying each obtained single-mode evaluation value by a sampling period adaptation coefficient to which the corresponding mode is attached in the step S22 so as to perform reliability modulation; And S25, carrying out weighted summation on each mode evaluation value subjected to credibility modulation according to the corresponding allocated mode weight so as to obtain a multi-mode evaluation score.
  3. 3. The method of claim 2, wherein the characteristic metrics of the myoelectric modality include a muscle contraction force achievement rate, a myoelectric activation degree score, and a muscle contraction stability score, wherein the characteristic metrics of the brain electric modality include an attention score and a pain correlation index, and wherein the characteristic metrics of the posture modality include a joint movement angle achievement rate, a movement stability score, and a movement trajectory deviation value.
  4. 4. The method according to claim 1, wherein the step S3 comprises: S31, based on the multi-mode evaluation score, matching a target rehabilitation action from a standard action library pre-stored with a plurality of standard rehabilitation training actions and corresponding biomechanical parameters thereof; s32, constructing at least one group of initial training parameters based on standard action parameters of a target rehabilitation action and combining individual parameters of a target object to form an initial individual rehabilitation treatment scheme; S33, in the implementation of the treatment scheme, myoelectric signals, electroencephalogram signals and posture data of the target object are acquired in real time, the real-time data are compared and checked with standard action parameters and expected representation values set based on initial training parameters, and the initial training parameters are dynamically adjusted according to the comparison result so as to optimize the individual rehabilitation treatment scheme.
  5. 5. The method according to claim 4, wherein the step of dynamically adjusting the initial training parameters in step S33 includes providing text, voice or visual guidance through real-time feedback and adjusting at least one of training intensity, difficulty of action or duration of training when the real-time comparison checks that at least one of muscle contraction force, joint movement angle or movement stability does not reach the expected representation value.
  6. 6. The method of claim 1, wherein the preprocessing comprises layered denoising and feature extraction of the electromyographic signals, attention concentration and pain correlation index assessment of the electromyographic signals, and joint key point identification and kinematic parameter extraction of the pose data.
  7. 7. The method of claim 6, wherein the steps of performing layered denoising processing and feature extraction on the electromyographic signals in the preprocessing include: Eliminating power frequency interference based on a notch filter with self-adaptive bandwidth; calibrating a signal baseline based on an algorithm of a signal time sequence gradient so as to eliminate low-frequency drift; separating the electromyographic signals and the composite electromagnetic interference based on an independent component analysis algorithm combined with dynamic sparse constraint, and carrying out signal reconstruction; performing feature verification on the myoelectric signals before and after denoising to ensure that the myoelectric key features exist; And extracting myoelectricity amplitude, integral myoelectricity and root mean square value in a dynamic time window based on the denoised myoelectricity signal.
  8. 8. The method according to claim 1, wherein the personalized adaptation parameters of the target object are obtained by collecting body state parameters and damage characteristic information of the target object, comparing the body state parameters and damage characteristic information with group evidence-based rehabilitation data in a database, dynamically calibrating training target parameters through an individual-group data comparison algorithm, and generating a training threshold range of the target object as the personalized adaptation parameters.
  9. 9. The method according to claim 1, further comprising a step S0 of determining a target muscle tissue by determining a type and a location of a damage to a muscle of the target subject before the step S1, wherein the target muscle tissue is a subject from which the myoelectric signal is collected.
  10. 10. A system for implementing the method according to any one of claims 1-9, comprising: a lesion diagnostic module configured for performing a lesion localization and type determination on muscle lesions of a target subject to determine target muscle tissue; A data acquisition processing module configured to include: the myoelectricity detection module is used for collecting myoelectricity signals of target muscle tissues; the electroencephalogram acquisition module is used for acquiring electroencephalogram signals of the target object; The gesture acquisition module is used for acquiring the human gesture of the target object; the preprocessing module is used for preprocessing the electromyographic signals, the electroencephalogram signals and the gesture data to obtain electromyographic characteristic data, electroencephalogram index data and gesture parameter data, so as to form multi-modal data; A fusion evaluation module configured to perform fusion evaluation based on the multimodal data to obtain a multimodal evaluation score; a protocol generation module configured to generate and execute an individual rehabilitation therapy protocol based on the multimodal assessment score and the personalized adaptation parameters of the target subject; The progress tracking and updating management module records the rehabilitation progress data of the target object and judges the rehabilitation state based on the rehabilitation progress data, when the rehabilitation state is detected to reach the preset threshold value, the data acquisition and processing module and the fusion evaluation module are triggered to work again so as to generate a new multi-mode evaluation score, and the individual rehabilitation treatment scheme is updated based on the newly generated multi-mode evaluation score.

Description

Muscle injury rehabilitation treatment method and system based on multi-mode data fusion Technical Field The invention relates to the crossing fields of rehabilitation medicine, biomedical engineering and computer technology, in particular to a muscle injury rehabilitation treatment method and system based on multi-mode data fusion. Background With the rapid development of medical cross technology, rehabilitation therapy is deeply transformed from a traditional and experience-based mode to a data-based, accurate and personalized direction. The brain-computer interface, myoelectricity detection, computer vision and other technologies provide a new tool for rehabilitation evaluation and intervention. However, the prior art still has remarkable limitation on the clinical application level, and the full play of the effect is restricted. First, the current technology focuses on single function rehabilitation scenarios or single modality data applications. For example, surface myoelectricity alone is used to assess muscle activation status, or motion capture system alone is used to assess joint mobility. Such single-dimensional assessment is difficult to comprehensively reflect the actual rehabilitation status of the patient. Rehabilitation of muscle injuries involves not only the restoration of physiological function of the muscle itself (e.g. strength, endurance), but also the reconstruction of motor control patterns (e.g. coordination, stability), and the regulation of neuropsychological factors (e.g. pain tolerance, training attention). Failure of any single modality to cover all of this, results in one-sided assessment, and treatment regimens formulated accordingly tend to be exclusive. Secondly, the prior art has insufficient capability in the collaborative application of multi-mode data, and especially lacks a mechanism for effectively fusing asynchronous and non-equal period multi-source data. The electromyographic signal sampling frequency is high (usually several kilohertz), the change is rapid, the change of the neuropsychological state reflected by the electroencephalogram signal is relatively slow, and the vision-based posture data is limited by the frame rate of a camera and the processing delay of an algorithm. These data are naturally unsynchronized and periodically inconsistent on the time axis. The simple time splicing or independent analysis and then merging can introduce time sequence mismatch, so that fusion evaluation results are distorted, and the cooperative state of a 'muscle-nerve-joint' system at the same moment can not be accurately reflected. How to realize high-precision time sequence alignment and credibility weighted fusion of multi-mode data is a key technical problem which cannot be well solved by the prior art. Furthermore, the individuation and dynamic adaptability of the existing rehabilitation system are generally insufficient. Many systems train based on fixed, standardized thresholds or protocols that fail to adequately account for individual differences in patients, such as lesion type and extent, physical characteristics, basal muscle strength, pain sensitivity, and the like. Meanwhile, once a treatment scheme is generated, the capability of dynamic adjustment based on real-time performance is often lacking, accurate guidance and parameter optimization cannot be given according to time conditions in the training process, so that the training efficiency is low, and secondary damage possibly caused by improper training is even caused. In addition, the adaptation disease range of the prior art is relatively narrow, and is mostly concentrated on neurogenic rehabilitation such as hemiplegia after cerebral apoplexy and spinal cord injury or conventional limb fracture postoperative rehabilitation, and the deep and accurate rehabilitation support for fields such as complicated skeletal muscle injury, ligament cartilage injury, muscle dysfunction accompanied by neurodegenerative diseases and the like is less, so that the increasingly diversified clinical rehabilitation requirements are difficult to meet. In view of the foregoing, there is a need for a muscle injury rehabilitation solution that can integrate multidimensional information, achieve accurate asynchronous fusion assessment, and provide full-cycle personalized dynamic therapy on the basis of the multidimensional information. Disclosure of Invention The invention aims to overcome the defects of the prior art and provides a multi-mode data fusion-based muscle injury rehabilitation method and system, which are characterized in that a full-closed-loop intelligent rehabilitation system of acquisition, evaluation, treatment and tracking is constructed, and the scientificalness, the accuracy and the intellectualization of muscle injury rehabilitation are realized through an innovative multi-mode asynchronous data fusion evaluation and individuation dynamic adaptation mechanism. The invention discloses a muscle injury rehabilita