I’m a newbie and trying to make a simple face-swapping program that utilizing a haar cascade algorithm to recognize the face and then by thresholding a grayscale of the face.
I’m inputting every pixel that matches the coordinates of a white pixel in the threshold image, but the value changes to 255.
I don’t really know why I’ve tried many different ways and am feeling kinda stuck.
Also, I’m thinking of buying the course off of their official website any recommendations?
It’s hard to summarize the critical points of face-swapping. May be this link which include many open-source program of face-swapping can help you.
Face detection is the first step of face-swapping and face alignment may be necessary. You can refer to the steps in former links.
1.Detect face.
2.Detect points on face.
3.Face triangulation.
4.Pise-wise affine warp.
5.Poisson blending.
6.Android or iOS integration.
Hope this can help you.
1 Like