I'm trying to find a solution for 2D face tracking (or some kind of motion tracking) on video input.
Is there a way to implement openCV face analysis in a shader ? or in a python script ?
Is it possible to pass some texture in argument to those ?
I am currently using a program called 'SplitCam' to create a virtual Direct Show signal, that can both be opened in Smode and in Max Msp. I guess there is at least 1 frame of latency.
Max Msp does the motion tracking with openCV and sends infomations back to Smode via OSC.
Do you have another lead or ideas ?