CS 194:-26 Image Manipulation and Computational Photography, Fall 2018

Project 4: Image Morphing

Jun Zhou, CS194-26-abm


By matching correspondences of two faces and triangulating the resulting points, we can perform an inverse affine warp transform on all the points in each respective triangle to a new subspace then interpolating between the pixels of the source and target image. By using this process with equal interpolation, we computed the mid-way image for these two faces.

The Morph Sequence

By using a sequential step sequence, we can warp by different parameters to perform the full morph between one face and another. At each iteration, the warp_frac and dissolve_frac will increase by the step size. By compiling all the resulting warped images, we can form the full morph sequence as a GIF.

The Mean Face of A Population

We compute the average shape of Danish face, then warp all faces into the shape. I show 3 examples of what some of them look like, and above is the average Danish face with all faces warped and averaged.


I found the average female Danish face and morphed my face into it for a caricature.

(BELLS & WHISTLES) Friends and Me