Project 4: Face Morphing

Background

For this project, we used Delaunay triangles and affine transformations to morph between two pictures.

Approach

The general approach was to first define corresponding set of points between to images. Find the Delaunay triangulation of the average shape of the points. Then, compute affine transformations of each Delaunay triangle to move from source image to target shape. Finally, we cross dissolved images to provide the correct colorization

Midway Point and Morphing

Here, I chose to warp my face with famous actor, Vin Diesal. The face position in each image is slightly off but the resulting warp seemed to be what we expect.

Me and Vin Diesal to average shape and GIF

Average Faces Dataset

Here, I used the Danes dataset to calculate average male and female faces.

Here are additional examples of some people warped to average structure