Currently we use opencv for extracting a feed from web cameras, but opencv is a big library only for that task. Also, as I understand it, it requires the python to be evaluated locally to access the hardware of the end-user.
My suggestion would be to use a library such as webrtc instead.
It uses the powers of HTML5, and could potentially connect to a end-user's hardware, even though python is evaluated remotely.
Currently we use opencv for extracting a feed from web cameras, but opencv is a big library only for that task. Also, as I understand it, it requires the python to be evaluated locally to access the hardware of the end-user.
My suggestion would be to use a library such as webrtc instead.
It uses the powers of HTML5, and could potentially connect to a end-user's hardware, even though python is evaluated remotely.
Also, it seems to be running more smoothly.
Best implementation for a fast integration seems to be https://github.com/maartenbreddels/ipywebrtc
However, it does not have an easy to use method for extracting the data.