Artificial Intelligence in Kinematics Analysis
Study Details
Study Description
Brief Summary
-
Establish data sets. The private data set includes relevant parameters including video of the subject's gait and standard methods for kinematic analysis;
-
Develop new models. Based on public and private data sets, the kinematic analysis model of human key point detection is further developed.
-
Test the new model. By comparing the parameters with the standard method, the accuracy of the model was verified, and the kinematics analysis model of artificial intelligence with accuracy above 98% was obtained
| Condition or Disease | Intervention/Treatment | Phase |
|---|---|---|
|
Detailed Description
Artificial intelligence human key point detection model mainly has traditional algorithm, "top-down" algorithm and "bottom-up" algorithm three methods, three methods have advantages. This project will comprehensively use the above three methods to conduct algorithm and parameter debugging in the public data set and test in the private data set, so as to obtain the most suitable human key point recognition method for gait analysis
Study Design
Arms and Interventions
| Arm | Intervention/Treatment |
|---|---|
| Normal subjects Gait analysis with artificial intelligence and traditional methods |
Device: Application Research of key points detection technology
Artificial intelligence human key point detection model mainly has traditional algorithm, "top-down" algorithm and "bottom-up" algorithm three methods, three methods have advantages. This project will comprehensively use the above three methods to conduct algorithm and parameter debugging in the public data set and test in the private data set, so as to obtain the most suitable human key point recognition method for gait analysis
|
| Subjects with abnormal gait Gait analysis with artificial intelligence and traditional methods |
Device: Application Research of key points detection technology
Artificial intelligence human key point detection model mainly has traditional algorithm, "top-down" algorithm and "bottom-up" algorithm three methods, three methods have advantages. This project will comprehensively use the above three methods to conduct algorithm and parameter debugging in the public data set and test in the private data set, so as to obtain the most suitable human key point recognition method for gait analysis
|
Outcome Measures
Primary Outcome Measures
- Gait related parameters [30mins]
Step frequency/pace/gait cycle/step length
Eligibility Criteria
Criteria
Inclusion Criteria:
-
- Abnormal gait.
-
Can walk 6m or more independently.
-
Older than 18.
Exclusion Criteria:
-
Fracture may be aggravated by walking in the acute stage or early postoperative stage. Have heart, lung, liver and kidney And other serious diseases, heart function grading greater than GRADE I (NYHA), respiratory failure and other symptoms and signs or Check the results.
-
The mental and psychological state cannot cooperate with the completion of the experiment.
-
High risk of falls (Berg score ≤20）
-
Gait kinematics analysis equipment cannot be used together.
Contacts and Locations
Locations
No locations specified.Sponsors and Collaborators
- Peking University Third Hospital
Investigators
None specified.Study Documents (Full-Text)
None provided.More Information
Publications
None provided.- M2021231