freemocap/freemocap
Freemocap
FreeMoCap is an open-source markerless motion capture system that reconstructs 3D human pose from video. It uses a multi-camera setup with ChArUco board calibration to accurately triangulate body landmarks, and it also supports single-camera recording for simpler captures. The system outputs skeleton joint data and generates interactive Jupyter notebooks for each recording, enabling users to explore and analyse motion data directly.
Built around hardware-synchronised video capture and MediaPipe-based 2D pose detection, FreeMoCap supports both calibrated multi-camera recording and real-time 2D/3D capture. It can also extract motion from existing video files. Recorded data is saved as organised files and can be exported to Blender for further editing or animation. The toolkit provides a complete pipeline from calibration through to analysis and export, all accessible without requiring physical markers.
Features
- Motion Capture - Captures 3D human motion from ordinary cameras or video files without markers for biomechanical analysis and animation.
- 3D Pose Reconstruction - Reconstructs 3D skeletal data from multiple calibrated camera views or existing video files.
- Camera Calibration - Uses a printed ChArUco board to calibrate multiple cameras for accurate 3D reconstruction.
- Multi-View Geometry Solvers - Triangulates corresponding 2D keypoints across calibrated camera views to compute 3D skeleton positions.
- Open Source Toolkits - An open-source toolkit that processes video to produce skeleton joint positions and analysis notebooks.
- Project-Specific Notebooks - Generates interactive Jupyter notebooks for each recording to explore and analyze motion data.
- Calibrated Multi-Camera Recordings - Synchronizes multiple cameras after calibration to capture 3D motion from multiple angles.
- Video-Based Reconstructions - Extracts 3D human motion from existing video files, a primary capability of the system.
- Motion Data Analysis - Generates interactive Jupyter notebooks enabling exploration and analysis of recorded motion data.
- Real-Time Motion Tracking - Captures body movements from connected cameras and outputs skeletal reconstructions in real time.
- Human Pose Detections - Detects 2D body landmarks from each camera view using a lightweight machine learning model.
- Multi-Camera Synchronization - Records from multiple cameras simultaneously using a hardware synchronization signal to align frames.
- File-Based Storage Systems - Organizes all recording data as local files (calibration, videos, keypoints, skeletons) for offline analysis.
- 3D Scene Exporters - Exports motion capture recordings as complete 3D scenes for editing and animation in Blender.
- Markerless Single-Camera Capture - Captures 3D skeletal motion from a single camera without markers or calibration.
