Advancing Monocular Video-Based Gait Analysis Using Motion Imitation with Physics-Based Simulation
Not on the map. Why?
Abstract (arXiv)
Gait analysis from videos obtained from a smartphone would open up many clinical opportunities for detecting and quantifying gait impairments. However, existing approaches for estimating gait parameters from videos can produce physically implausible results. To overcome this, we train a policy using reinforcement learning to control a physics simulation of human movement to replicate the movement seen in video. This forces the inferred movements to be physically plausible, while improving the accuracy of the inferred step length and walking velocity.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 21 of its 32 references that are in the map. First posted on arXiv 20 Feb 2024.
Builds on
21 of 32 references.
References · 21
- 2014Adam: A Method for Stochastic Optimizationcited by 36,960
- 2017Proximal Policy Optimization Algorithmscited by 14,441
- 2019AMASS: Archive of Motion Capture as Surface Shapescited by 1,208
- 2021Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learningcited by 882
- 2023Eureka: Human-Level Reward Design via Coding Large Language Modelscited by 542
- 2021Brax -- A Differentiable Physics Engine for Large Scale Rigid Body Simulationcited by 375
- 2023Perpetual Humanoid Control for Real-time Simulated Avatarscited by 198
- 2021SimPoE: Simulated Character Control for 3D Human Pose Estimationcited by 111
- 2022Accelerated Policy Learning with Parallel Differentiable Simulationcited by 92
- 2022MyoSuite -- A contact-rich simulation suite for musculoskeletal motor controlcited by 80
- 2023Reaching the Limit in Autonomous Racing: Optimal Control versus Reinforcement Learningcited by 72
- 2020Residual Force Control for Agile Human Behavior Imitation and Extended Motion Synthesiscited by 63
9 more
- 2020MoVi: A Large Multipurpose Motion and Video Datasetcited by 58
- 2022Learning 3D Human Pose Estimation from Dozens of Datasets using a Geometry-Aware Autoencoder to Bridge Between Skeleton Formatscited by 50
- 2023LocoMuJoCo: A Comprehensive Imitation Learning Benchmark for Locomotioncited by 31
- 2022PoseTriplet: Co-evolving 3D Human Pose Estimation, Imitation, and Hallucination under Self-supervisioncited by 25
- 2023Improved Trajectory Reconstruction for Markerless Pose Estimationcited by 12
- 2022PosePipe: Open-Source Human Pose Estimation Pipeline for Clinical Researchcited by 10
- 2023Markerless Motion Capture and Biomechanical Analysis Pipelinecited by 10
- 2022Transforming Gait: Video-Based Spatiotemporal Gait Analysiscited by 8
- 2023Self-Supervised Learning of Gait-Based Biomarkerscited by 5
Built on by
Citing papers per year: 2024: 1; 2025: 1.
2 papers.