CS281A Project: Nonlinear Dimensionality Reduction on Human Facial Expressions
SVM HeaderParse 0.2
AUTHOR NAME
Ryan White
SVM HeaderParse 0.2
ABSTRACT
In this paper we explore the utility of nonlinear dimensionality reduction techniques in the realm of facial expression analysis. First, we test the ability of nonlinear techniques to describe the higher nonlinear nature of human facial expressions. We exploit the data-driven model of an embedding to create novel facial expressions. Finally, we composite the facial expressions back on the face. 1