A fully neural approach to color facial image recognition
Victor-Emil Neagoe, Alexandru-Cristian Mugioiu · World Automation Congress · 2008
The paper proposes a neural network cascade for color facial image recognition consisting of three processing stages: a) conversion of the standard RGB color space into the 3D uncorrelated color space (UCS) using a Hebbian neural network (HNN), that implements the model of Principal Component Analysis (PCA) in the color space; b) fusion of the three UCS components; c) the neural classifier called Concurrent Self-Organizing Maps (CSOM). The system is experimented for a set of 906 color facial images of 151 subjects selected from the Essex University database. The very good experimental results for color face recognition are given.