Hand pose retargeting system for Wuji Hand. High-precision retargeting based on adaptive analytical optimization, supporting Apple Vision Pro hand tracking for real-time teleoperation.
retargeting_new_performance.mp4
Get started with Quick Start. For detailed documentation, please refer to Retargeting Tutorial on Wuji Docs Center.
├── wuji_retargeting/
│ ├── opt/
│ │ └── ...
│ └── wuji_hand_description/
│ └── ...
├── example/
│ ├── input_devices/
│ │ └── ...
│ ├── config/
│ │ └── ...
│ ├── data/
│ │ └── ...
│ └── utils/
│ └── ...
├── requirements.txt
└── README.md
| Directory | Description |
|---|---|
wuji_retargeting/ |
Core package containing retargeter interface, optimizer modules, kinematics, and coordinate transforms |
wuji_retargeting/opt/ |
Optimizer implementations including adaptive analytical optimizer |
wuji_retargeting/wuji_hand_description/ |
URDF and mesh submodule for Wuji Hand |
example/ |
Demonstration scripts for simulation and hardware control |
example/input_devices/ |
Input device modules (Vision Pro, MediaPipe replay) |
example/config/ |
YAML configuration files |
example/data/ |
Sample recording data |
git clone --recurse-submodules https://github.com/wuji-technology/wuji-retargeting.git
cd wuji-retargeting
pip install -r requirements.txt
pip install -e .cd example
# Simulation - replay recording
python teleop_sim.py --play data/avp1.pkl --hand left
# Real hardware
python teleop_real.py --play data/avp1.pkl --hand rightIf you find this project useful, please consider citing:
@software{wuji2025retargeting,
title={WujiHand Retargeting},
author={Guanqi He and Wentao Zhang},
year={2025},
url={https://github.com/wuji-technology/wuji_retargeting},
note={* Equal contribution}
}This project is built upon several excellent open-source projects:
- MuJoCo for physics simulation
- dex-retargeting for hand retargeting algorithms
- DexPilot for vision-based teleoperation insights
- VisionProTeleop for Apple Vision Pro streaming
For any questions, please contact support@wuji.tech.