Face Morphing and Modelling a Photo Collection
Sean Chen
Computing the "Mid-way Face"
Input A
Input B
Halfway face
The Morph Sequence
The "Mean face" of a population
Average face of Danish Computer Scientists
Subject 8 morphed to average face geometry
Subject 10 morphed to average face geometry
Subject 20 morphed to average face geometry
Subject 30 morphed to average face geometry
Subject 35 morphed to average face geometry
My face warped to the average geometry
Average Danish face warped to my geometry
Caricatures: Extrapolating from the mean
Extrapolating to have more of my facial geometry
Half Dane, half me
Extrapolating with extra danish facial geometry
Bells and Whistles
Average chinese woman
Average chinese man
Using the faces of the average chinese man and woman, I represented their facial geometry as vectors. Then, I subtracted the facial geometry of the men from the women, and then added that vector to my facial geometry. This results in a more feminine face. We can do this using only shape, only color. or both.
Shape only
Using both shape and color
Using my facial geometry, linear interpolation of colors