This is a ROS package designed for head pose imitation on the Surena V humanoid robot.
To estimate human head pose, we utilized the "Dense Head Pose Estimation" project. The obtained head pose information is then mapped onto our robot for imitation purposes.
-
The configuration files for the face detection and facial landmarks models should be placed in the
config/weights
directory. -
You can adjust the thresholds and parameters for face detection and facial landmarks in the
main.py
file. -
The mapping from detected angles space to motor commands space can be modified by updating the constants in the
main.py
file.
Contributions are welcome! If you find any issues or have suggestions for improvements, please open an issue or submit a pull request.
- Adding simulation support