In this assignment, I morph the shapes and colors of human faces to create fascinating results!
Part 1: Defining Correspondences
In this portion, I wrote a python program to take in two images and help the user choose corresponding points in the images. These points were input into the
code which morphed images. Additionally, I obtained a triangulation for both images using Delaunay triangulation.
Part 2: Computing the "Mid-way Face"
Here, I computed the midway object of these two images by computing the average shape of both images, warping both images to the average shape, and averaging the color of the
images together.
Shubha's Original Image
Anmol's Original Image
Shubha warped to Average
Anmol warped to Average
Average Shubha and Anmol
Part 3: The Morph Sequence
In this part of the project, I used linear interpolation to create a gif which smoothly transitions between two images. First, the images are warped to the
average shape. Then, I cross dissolve the colors.
Part 4: The "Mean Face" of a Population
I warped and combined all the images from the database of danish computer scientists to get the average danish computer scientist here!
Here were some of the Danish people morphed into the average face.
Danish Computer Scientist 1
Danish Computer Scientist 2
Danish Computer Scientist 3
Here I've warped my face to that of the average Danish computer scientist and the average Danish computer scientist to my face!
Me warped into the Danish average shape
Danish average shape warped to me
Part 5: Caricatures - Extrapolating From the Mean
Here's a charicature of me created by extrapolating my face from the face using the face of the average Danish computer scientist.
Part 6: (Bells and Whistles) Changing Gender
Here, I've turned my good friend Anmol into a girl by averaging his face with that of the average Indian girl.
Here were some of the Danish people morphed into the average face.
