Gait Analysis based on Azure Kinect 3D Human Skeleton

作者：Yan Zeng, Liufan Wu, Dong‐Lin Xie · 发表于：2021 International Conference on Computer Information Science and Artificial Intelligence (CISAI) · 年份：2021 · DOI：10.1109/cisai54367.2021.00212 · 被引用次数：8 · 研究领域：Gait Recognition and Analysis、Human Pose and Action Recognition、Diabetic Foot Ulcer Assessment and Management

Azure Kinect camera system can avoid the limitations of high cost, large system volume and cumbersome acquisition process of traditional gait capture system to collect human motion trajectory. However, the validity of gait feature extraction on recorded data has not been determined. In order to verify this, we collected the motion trajectory on 10 healthy individuals with Azure Kinect, extracted the gait features based on the obtained 3D human skeletons, and compared with Vicon data, which is regarded as the golden standard. The experimental results show that Azure Kinect can perform effective gait feature extractions, and verify the feasibility for gait analysis.