Detect face and hand landmarks using Python, Mediapipe and OpenCV

A Flask web app to switch on live video feed using webcam and detecting face and hand landmarks in real time.

How to start Flask app ?

Start virtual env

From cloned directory go to python_env and run

source ./bin/activate

Run app

python app.py

Once you start an app, you can paste address given by app in your browser and run http://

:/ to check if it's working fine -> prints ok http://:/track to start live video feed and start detecting face and hand landmarks

To deploy in Docker container

Build Docker Image

docker  build  -t  <image_name>:<tag>  .

Start docker container

docker  container  run  -d   -p  5000:5000   <image_name>:<tag>

You might need to enable your webcam access for Docker Container, add below flag while starting docker container to enable webcam access inside container.

--device=/dev/video0:/dev/video0

https://stackoverflow.com/questions/44852484/access-webcam-using-opencv-python-in-docker

Feel free to clone and improve it further and raise ticket if required.