This project was to morph an image of one face into another face. To do this I set a standard set of control points on each face. There were 43 points on the face, and I also include the 4 cornes and midpoints of the edges for a total of 51 control points.
![]() |
![]() |
I computed the average of the two triangulations and used affine transformations to warp the triangles from each image into the average triangulation.
![]() |
![]() |
After both images were normalized to the same shape, I just dissolved the two images to create a halfway warp between the two faces. Here is my result of warping halfway:
The same technique can be used to create a face that is a certain percent of each image, so by slowly increasing this percent, you can create a morph between the two faces. The individual frames for this morph can be seen here.
Class Average
An average face can be computed by using the triangulations of each face to find an average shape. Then each face can be morphed into this shaped an averaged together. Here are my results:
![]() |
![]() |