Automatic face replacement in images and videos
Authors:
The goal of the project is to automatically detect and seamlessly replace all the faces in an image with a reference face. The pipeline for face replacement in images is then extended to videos. This was the final project for CIS 581 Computer vision and computational photography done in 2014.
Results of face replacement in video: https://www.youtube.com/watch?v=DwmR5lUuXT4
Run main.m
main.m
- wrapper file which calls face_replace.m
with the input image/video and the face to replace
face_replace.m
- Replaces the face(s) in image/video and displays it. (The video is saved as result.avi
)
extafacedescs.m
- wrapper to call 3rd party lib to extract facial feature points
getConvHull.m
- our function, calculates the angle at which the face is inclined to with the vertical and a label for pose (left, frontal, right) using geometry. It then also estimates some points on the face which form a convex hull without cutting out the person’s eyebrow
ref_images.mat
- library of reference faces in different poses