Closed chuigda closed 1 year ago
VTS's face capture is based on OpenSeeFace. Maybe we can directly utilize OSF, bypassing the VTS's rendering. And after that, we won't need to include any proprietary software in our release.
OSF requires a little complex setup, and comes without some utilities (meaning that we have to implement our own calibration things or so). Let's just stick to VTS by this time.
Well done @CousinZe
找了一些可以用来参考的 repository
此外可以考虑下是不是不用
solvePNP
也能完成任务