This project ‘face morphing’ is divided into 7 parts to complete a smooth transfer video from two faces;

1. ‘dlib’ face landmark detector is used to detect 68 landmarks on 2 images, as 68 landmarks are not covering all the face, 20 landmarks are manually added, including 16 landmarks on the face, while 4 points locates the background, so that the background could also transfer smoothly;
2. the standard Delaunay Triangulation is used to visualize triangulation of vertices(landmarks);
3. get the interpolating between image1 and image2 by applying x int = (x img1 + ximg2)/2 y int = (y img1 + yimg2)/2
4. Getting parameters of the affine warp when vertices of two triangle were given, and between every pairs of corresponding triangles, Affine warp was estimated;
5. Find the all coordinates which are inside the given triangle, and then a labeled image was created by dividing face into parts;
6. Get the in - between image for a given ‘w’
7. Create the transfer video;