Recognizing facial expressions : Computational models and humans

Abstract

This paper discusses various biologically plausible computational models that recognize human facial expression and analyze them. Identifying facial expressions is a non trivial task for a human and is a key part of social interactions. However, it is not as simple as that for a computational system. Here we analyze six different universally accepted facial expressions for analysis with the aid of six biologically plausible computational models. There have been a limited number of studies comparing the performance of human subjects with computational models for facial expression recognition. This paper does a genuine attempt in making this comparison.