Method for performing motion analysis on three-dimensional image pairing under assistance of visual motion capture
Through the combination of visual motion capture and CT scanning technology, the problem of insufficient accuracy of two-dimensional image data and soft tissue hard tissue registration in traditional rehabilitation analysis methods is solved, and high-precision three-dimensional motion analysis is achieved, which improves the accuracy and personality of rehabilitation assessment. Serving services.
Patent Information
- Application Number
- CN202411156916.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-22
- Publication Date
- 2024-11-22
AI Technical Summary
The traditional rehabilitation analysis method relies on two-dimensional image data and manual labeling, with limited accuracy and cumbersome operation, which cannot meet the needs of high-precision and real-time analysis. The existing analysis method based on visual motion capture has limitations in data pairing and processing. Especially when it comes to registration between soft and hard tissues.
Using visual motion capture assisted method, three-dimensional motion data is collected through optical binocular stereo tracking cameras, combined with CT scanning technology to generate a three-dimensional image model, data pairing is used using a registration algorithm based on feature point matching, and a machine learning algorithm is used to identify motion patterns. , extract key motion parameters for analysis.
High-precision motion analysis is achieved, deformation registration problems between soft and hard tissues are solved, the accuracy of rehabilitation assessment and personalized services are improved, and the patients' rehabilitation effect and quality of life are improved.
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of motion analysis, and in particular to a method for motion analysis using three-dimensional image pairing assisted by visual motion capture. Background Art
[0002] Traditional rehabilitation analysis methods mainly rely on two-dimensional image data or manual marking. This method not only has limited accuracy, but also is cumbersome to operate and cannot meet the needs of high-precision and real-time analysis. In recent years, the development of visual motion capture technology has provided a new approach for rehabilitation. By capturing and analyzing the patient's three-dimensional motion data, more accurate and comprehensive motion analysis can be achieved. However, existing analysis methods based on visual motion capture still have certain limitations in data pairing and processing, especially in the registration problem between soft tissue and hard tissue. Summary of the invention
[0003] In order to make up for the defects of the prior art, the present invention provides a method for motion analysis using three-dimensional image pairing assisted by visual motion capture.
[0004] The present invention is achieved through the following technical solutions: A method for motion analysis using three-dimensional image pairing assisted by visual motion capture comprises the following steps: 1) Collect the patient’s three-dimensional motion data; 2) Preprocessing the collected three-dimensional motion data, including noise filtering, data smoothing and coordinate confirmation; 3) Generate a three-dimensional image model of the patient using CT scanning technology; 4) Use a registration algorithm to pair the preprocessed 3D motion data with the 3D image model; 5) performing lower limb motion analysis based on the pairing results, wherein the lower limb motion analysis includes judging human posture, tracking motion trajectories, extracting motion parameters, and recognizing motion patterns; 6) Display the analysis results through visualization tools and provide rehabilitation assessment and suggestions.
[0005] The method for motion analysis using three-dimensional image pairing assisted by visual motion capture comprises the following steps: 1) Use an optical binocular stereo tracking camera to collect the patient's 3D motion data: The patient wears a shaping sock with a positioning ball, ensuring that the positioning ball is installed at each key joint; the patient's 3D motion data is generated through a multi-view reconstruction algorithm; 2) Preprocessing the collected 3D motion data, including using filters to filter noise and using spline interpolation to smooth data and confirm coordinates to ensure data accuracy; 3) Generate a three-dimensional image model of the patient using CT scanning technology; 4) Using a registration algorithm based on feature point matching, the pre-processed 3D motion data is paired with the CT image model; the feature points include joint points and limb nodes. By calculating the distance and angle relationship between the feature points, the deformation registration problem between soft tissue and hard tissue is solved to ensure the pairing accuracy; 5) Track the patient's lower limbs based on the pairing results and extract key motion parameters, including speed, acceleration, and angular velocity; use machine learning algorithms to identify motion patterns, classify and evaluate the patient's movements, predict disease progression, and assist in formulating rehabilitation treatment strategies; 6) Display the analysis results through visualization tools, provide rehabilitation assessment and improvement suggestions, and help patients improve rehabilitation effects and quality of life.
[0006] The beneficial effects of the present invention are: The present invention captures and analyzes human motions through a camera, aligns with CT three-dimensional images, scientifically and quantitatively analyzes human motion characteristics, realizes analysis and comparison under different postures, and then realizes the transformation from a method based on physical examination and personal diagnosis and treatment experience to a human body and motion analysis method based on computer vision through the correspondence between various actions and anatomical structures under digital simulation, thereby providing better personalized services for the promotion of sports concepts and the improvement of early prevention and prediction levels of patients. The present invention can realize high-precision medical rehabilitation analysis and improve the efficiency and effect of patients' rehabilitation level detection in outpatient clinics. The method and system can not only be applied to patients' rehabilitation and disease prediction, but also can be widely used in medical research, rehabilitation and health management and other fields. DETAILED DESCRIPTION Example
[0007] The method for motion analysis using three-dimensional image pairing assisted by visual motion capture comprises the following steps: 1) Use an optical binocular stereo tracking camera (such as NDI equipment) to collect the patient's 3D motion data: the patient wears a shaping sock with a positioning ball, ensuring that the positioning ball is installed at each key joint; the patient's 3D motion data is generated through a multi-view reconstruction algorithm; 2) Preprocessing the collected 3D motion data, including using filters to filter noise and using spline interpolation to smooth data and confirm coordinates, to ensure the accuracy and reliability of the data; 3) Use CT scanning technology to generate a three-dimensional image model of the patient to ensure that the three-dimensional image model can truly reflect the patient's actual body structure; 4) A registration algorithm based on feature point matching is used to match the preprocessed 3D motion data with the CT image model; the feature points include joint points and limb nodes. By calculating the distance and angle relationship between the feature points, the deformation registration problem between soft tissue and hard tissue is solved to ensure the accuracy of the pairing; 5) Track the patient’s lower limbs based on the pairing results and extract key motion parameters, including speed, acceleration, and angular velocity; use machine learning algorithms to identify motion patterns, classify and evaluate the patient’s movements, predict disease progression, and assist in developing rehabilitation treatment strategies; 6) The analysis results are presented to doctors and patients through visualization tools, providing rehabilitation assessment and improvement suggestions, assisting in formulating treatment strategies, predicting disease progression, tracking patient recovery, and helping patients improve rehabilitation effects and quality of life.
[0008] Except for the technical features described in the specification, the other technical features are known technologies to those skilled in the art.
Claims
1. A method for motion analysis using three-dimensional image pairing assisted by visual motion capture, characterized in that: The steps include: 1) Collect the patient’s three-dimensional motion data; 2) Preprocessing the collected three-dimensional motion data, including noise filtering, data smoothing and coordinate confirmation; 3) Generate a three-dimensional image model of the patient using CT scanning technology; 4) Use a registration algorithm to pair the preprocessed 3D motion data with the 3D image model; 5) performing lower limb motion analysis based on the pairing results, wherein the lower limb motion analysis includes judging human posture, tracking motion trajectories, extracting motion parameters, and recognizing motion patterns; 6) Display the analysis results through visualization tools and provide rehabilitation assessment and suggestions.
2. The method for motion analysis using three-dimensional image pairing assisted by visual motion capture according to claim 1, characterized in that: The following steps are involved: 1) Use an optical binocular stereo tracking camera to collect the patient's 3D motion data: The patient wears a shaping sock with a positioning ball, ensuring that the positioning ball is installed at each key joint; the patient's 3D motion data is generated through a multi-view reconstruction algorithm; 2) Preprocessing the collected 3D motion data, including using filters to filter noise and using spline interpolation to smooth data and confirm coordinates to ensure data accuracy; 3) Generate a three-dimensional image model of the patient using CT scanning technology; 4) Using a registration algorithm based on feature point matching, the pre-processed 3D motion data is paired with the CT image model; the feature points include joint points and limb nodes. By calculating the distance and angle relationship between the feature points, the deformation registration problem between soft tissue and hard tissue is solved to ensure the pairing accuracy; 5) Tracking the patient's lower limbs based on the pairing results and extracting key motion parameters, including velocity, acceleration, and angular velocity; Use machine learning algorithms to identify movement patterns, classify and evaluate patients’ movements, predict disease progression, and assist in developing rehabilitation treatment strategies; 6) Display the analysis results through visualization tools, provide rehabilitation assessment and improvement suggestions, and help patients improve rehabilitation effects and quality of life.
Citation Information
Cited By
Multi-modal information-based motion function rehabilitation state detection method
CN119523473A
