This application explores the functionality of some of Google's Mediapipe Machine Learning solutions, viz:
- Hand Tracking
- Pose Estimation
- Face Detection
- Face Mesh
StreamLit is used to create the Web Graphical User Interface (GUI). Streamlit is a cool way to turn data scripts into shareable web apps in minutes, all in Python. You can check out the launch blog post for more information.
This web app was inspired by the awesome YouTube tutorials created by:
- Murtaza Hassan: Advanced Computer Vision with Python - Full Course
- Augmented Startups: StreamLit Computer Vision User Interface Course | MediaPipe OpenCV Python (2021)
Do check out their channels and websites for more informative and exciting Machine Learning & Computer Vision tutorials.
The app was been deployed on Streamlit. You can check it out here.
There have been a couple of issues running video sources on the online shared app:
- There is a considerable lag when running the Mediapipe modules on videos with one viewer/user.
- If multiple viewers/users attempt to use video inputs at the same time, the app becomes totally unresponsive and needs to be rebooted.
I suspect that both issues are related but I have been unable to get them fixed. Kindly reach out if you have any ideas/suggestions on how I can go about resolving these issues.
If you want to test out video sources, you can clone the source repository, install the requirements and run streamlitMediapipe.py.