Combining Deep Learning Technologies with Multi-Level Gabor Features for Facial Recognition in Biometric Automated Systems

摘要

Face recognition is one of the most important abilities that humans use in their daily lives. It represents a natural, robust and non-intrusive manner for identifying individuals. However, it is also a very challenging problem in the field of computer vision and pattern recognition. A good face recognition algorithm should be able to automatically detect and recognize a face in an image, regardless of lightning, expression, illumination and pose. In this paper, we present a novel approach for the face model representation and matching issues in face recognition. Our approach is based on multi-level Gabor features and Deep Learning techniques. In the experiments presented in this paper, ORL, Caltech, Yale and Yale B databases were used in order to obtain the face recognition rate. The results show that the new face recognition algorithm outperforms the conventional methods such as global Gabor face recognition based on PCA in terms of recognition rate.