SynthGait19K dataset improves gait parameter estimation from video
SynthGait-19K: A Physically Grounded Synthetic Video Dataset for Gait Parameter Estimation
Computer Vision and Pattern Recognition
Summary
Measuring how people walk using just regular video is hard because existing data is limited and not very diverse. The authors created SynthGait-19K, a large synthetic video dataset of people walking, based on detailed motion capture data, which includes important walking measurements. They built a system to make realistic walking videos from this data that can be viewed from different angles and settings. Their tests show that training computer models with this synthetic data helps predict walking parameters on real videos better, especially for some types of walking features.
What this means in practice
- •For physical therapy teams: Estimate key gait parameters from ordinary videos to support scalable patient mobility assessments without specialized equipment.
- •For mobile health app developers: Enhance video-based gait tracking features in health apps by training models with diverse synthetic walking videos for better real-world performance.$Commercial implications: Enables sale of more accurate gait monitoring features to consumers and healthcare providers using smartphones.
Authors
Soroush Mehraban, Xin Lei Lin, Vida Adeli, Majid Mirmehdi, Amirhossein Dadashzadeh, Clint Hansen, Andrea Iaboni, Babak Taati
Abstract
Accurate estimation of clinically meaningful gait parameters from monocular video is important for scalable mobility assessment, yet progress is limited by the small scale, restricted viewpoints, and limited visual diversity of existing datasets. We introduce SynthGait-19k, a physically grounded synthetic video dataset containing 19,272 walking videos derived from 6,427 MoCap sequences across 437 subjects, with paired SMPL motion and annotations for six gait parameters. To construct the dataset, we develop Gait2Vid, which unifies heterogeneous MoCap recordings through SMPL and synthesizes diverse RGB walking videos under controllable viewpoints and scene appearances. We assess the generated videos for consistency with their conditioning gait kinematics and validate extracted gait events against force-platform measurements. Using SynthGait-19K, we benchmark direct RGB, pose-based, biomechanical, and human-mesh-recovery approaches and analyze viewpoint, training-data scale, and synthetic-to-real domain shift. We also introduce GaitXFormer as a direct RGB reference model for estimating gait parameters. Synthetic supervision transfers effectively to real videos across both GaitXFormer and a pose-based architecture, demonstrating utility across different representations. We further find that spatial gait parameters are more sensitive to visual domain shift and that improved HMR reconstruction alone does not necessarily translate to improved downstream gait estimation.