Using moment invariants and HMM in facial expression recognition

Moment invariants are invariant under shifting, scaling and rotation. They are widely used in pattern recognition because of their discrimination powe r and robustness. HMM method is a natural and highly reliable way of recogn ition. In this paper, we have proposed a method of using moment invariants as features and HMM as recognition method in facial expression recognition. Sequences of four universal expressions, i.e. anger, disgust, happiness an d surprise, are recognized. We were able to attain an accuracy as high as 9 6.77%. (C) 2002 Elsevier Science B.V, All rights reserved.