Face recognition using principle components and linear discriminant analysis
ISPRA'09 Proceedings of the 8th WSEAS international conference on Signal processing, robotics and automation
| Hi-index | 0.00 |
Recent study demonstrates that color information makes contribution and enhances robustness in face recognition. Yet, most of the approaches to face recognition still use gray-scale images, mainly due to the extra processing and storage costs of the larger input size for color images. Multi-layer neural networks (MLNs) have been widely used in face recognition applications and yielded highly competitive results. This paper proposes a new method that makes a full use of color information without noteworthy extra processing cost, and demonstrates that the neural network in the proposed method is compatible with the existing MLN training algorithms based on error back propagation (EBP). Experimental results verify the superiority of the proposed method in face recognition.