Skip to main content
. 2024 Nov 26;3(11):e0000668. doi: 10.1371/journal.pdig.0000668

Fig 4. Deep learning-based locomotive syndrome prediction method.

Fig 4

a. Steps involved in deep learning-based method for locomotive syndrome prediction. Step 1: Video recording of the subject. Step 2: Pose estimation conducted using OpenPose. Step 3: Development of the LS prediction model utilizing MS-G3D. Step 4: Final prediction of Locomotive Syndrome. LS stands for Locomotive Syndrome. b. Skeleton model generated by OpenPose. Depicts the 2D coordinates for 25 key body points as identified by the OpenPose framework. c. Spatial-Temporal GCN-Based LS Prediction Model Utilizing MS-G3D. Diagram of the Spatial/Temporal Graph Convolutional Network (GCN) component. d. Spatial-Temporal GCN-Based LS Prediction Model Utilizing MS-G3D. Enhanced Spatial-Temporal GCN architecture incorporating skip connections.