WebRTC video chat and Pose Detection

I want to run TensorFlow pose detection on the WebRTC video chat web through Firebase.
However, the video brought through the server cannot be pose-detected, and errors continue to occur that there is no ‘onloded data’.
Is there a good way to solve it?

Do you have a link to the live demo or provide the exact errors. It is hard to debug without these details. My guess however is that you need to:

  1. Serve the webpage over HTTPS if trying to use webcam - else access will be denied.

  2. Ensure CORS headers are set for video from remote sources or domains else your HTML Canvas will be tainted and not accessible in JavaScript. Please check resources on CORS if this is the issue:

If you want a website that allows you to host simple websites for free on https and try out HTML/CSS/JS ideas check out Glitch.com. We even have a presence on there where you can fork our already working code as a starting point eg getting webcam imagery etc and classifying it.