FreeMoCap

FreeMoCap is a free, open-source motion capture system developed by researchers at the University of Texas at Austin. It provides a hardware- and software-agnostic, minimal-cost, research-grade motion capture platform that works with standard consumer webcams. The project is licensed under AGPL and hosted on GitHub at github.com/freemocap/freemocap.

Traditional motion capture systems can cost tens of thousands of dollars and require specialized hardware and dedicated studio space. FreeMoCap democratizes motion capture by enabling anyone with a few USB webcams to capture high-quality 3D motion data for scientific research, biomechanics education, animation, and clinical applications.

Quick Install via pip (Python 3.10–3.12):

pip install freemocap

After installation, launch the GUI from the command line:

freemocap

A graphical interface will appear, guiding you through camera calibration, recording, and data export.

Install from Source

Clone the repository and install in development mode:

git clone https://github.com/freemocap/freemocap
cd freemocap
pip install -e .

FreeMoCap uses a multi-camera setup with consumer webcams (typically 2–4 cameras) to capture synchronized video footage. The system then processes this footage through:

1. Camera Calibration — Uses a checkerboard pattern to calibrate camera intrinsics and extrinsics, establishing the spatial relationship between cameras
2. 2D Pose Estimation — Applies deep learning-based body tracking (MediaPipe, OpenPose, or DANNCE) to detect 2D joint positions from each camera view
3. 3D Triangulation — Reconstructs 3D joint positions using multi-view geometry and triangulation algorithms
4. Post-Processing — Applies filtering, gap-filling, and smoothing to produce clean motion data

• Hardware Agnostic — Works with any USB webcam; no specialized motion capture hardware required
• Minimal Cost — The entire system can be set up for under $100 using consumer-grade cameras
• Research Grade — Validated against gold-standard motion capture systems for biomechanics research
• Multi-Person Tracking — Supports tracking multiple subjects simultaneously in the same space
• Real-Time Processing — Optional real-time 2D pose estimation with post-hoc 3D reconstruction
• Open Source — AGPL-licensed with full source code available; contributions welcome

FreeMoCap exports motion capture data in multiple formats suitable for analysis, visualization, and integration with other tools:

Format	Description
.trc	Motion analysis format compatible with biomechanics software
.c3d	Standard biomechanics data format for gait analysis
.csv	Tabular data with joint positions and angles over time
.blend	Blender scene files with animated skeleton for visualization

FreeMoCap has been used in a wide range of research, education, and creative applications:

• Biomechanics Research — Gait analysis, movement disorders, rehabilitation assessment
• Sports Science — Athletic performance analysis, technique optimization
• Animation & VFX — Character animation reference and motion data for 3D pipelines
• Clinical Applications — Remote patient monitoring, physical therapy progress tracking
• Education — Teaching biomechanics, computer vision, and motion analysis concepts
• Robotics — Motion data for imitation learning and human-robot interaction research

FreeMoCap has an active research community and comprehensive documentation:

• Documentation — Full documentation and beginner tutorials at freemocap.github.io/documentation
• Discord Community — Active community for support, collaboration, and sharing projects
• GitHub — Open-source codebase with contribution guidelines at github.com/freemocap/freemocap
• Academic Citation — The project is citable via Zenodo with DOI: 10.5281/zenodo.7233714