OpenCV.js facial landmarks + human skeleton tracking

asked 2018-03-02 05:06:39 -0600

kavikode gravatar image

Hi Everyone,

I am trying to use OpenCV.js for human skeleton tracking of the arms and legs and eye blink and smile detection.

I am not sure how to extract the facial landmarks with OpenCV.js.

Can someone please help me by providing an example guidance on the logical steps to follow to make this possible on the web browser?

It sounds simple but I'm not sure how to begin.

Please help. I appreciate everyone's help.

edit retag flag offensive close merge delete


"It sounds simple " -- it isn't, clearly. all of opencv.js, skeleton(pose), facial landmarks are fairly recent additions, bleeding edge.

but your question is about 4 tasks at once, can you rather narrow it down to one problem at a time ?

I am not sure how to extract the facial landmarks with OpenCV.js.

this might not be possible at all from js, for now.

berak gravatar imageberak ( 2018-03-02 05:11:48 -0600 )edit

understand, thanks.

What is the best approach to handle skeleton (pose) tracking from the webcam stream of a web browser?

How do I extract the facial landmarks with OpenCV.js?

kavikode gravatar imagekavikode ( 2018-03-02 05:27:00 -0600 )edit

there is a python sample for openpose and a js sample for dnn face recognition , you'll have to adapt those, somehow.

extracting landmarks is not possible at all from js.

there's also a haarcascade to detect smiles here , but it's so-la-la.

berak gravatar imageberak ( 2018-03-02 05:32:31 -0600 )edit