Expressions Invarient Face Recognition

Abstract

Variation of Facial expression is one of the most challenging factors in face recognition which significantly degrades the performance of the face recognition system. This paper presents a mechanism for reducing the high false positive rate in face recognition due to facial expression by using a novel Gaussian blurring and decimation algorithms. Extensive experimentation on complex images with variant facial expressions from the ORL dataset shows that removing the higher frequencies in the pre-processing step enhances the performance of eigenfaces algorithm by a significant amount i.e. from 93.75% to 96.5%