Color by Correlation: A Simple, Unifying Framework for Color Constancy
IEEE Transactions on Pattern Analysis and Machine Intelligence
Mean Shift: A Robust Approach Toward Feature Space Analysis
IEEE Transactions on Pattern Analysis and Machine Intelligence
Digital Image Processing
Concurrent Self-Organizing Maps for Pattern Classification
ICCI '02 Proceedings of the 1st IEEE International Conference on Cognitive Informatics
A Neural Approach to Compression of Hyperspectral Remote Sensing Imagery
Proceedings of the International Conference, 7th Fuzzy Days on Computational Intelligence, Theory and Applications
Optimization of color conversion for face recognition
EURASIP Journal on Applied Signal Processing
ICA color space for pattern recognition
IEEE Transactions on Neural Networks
Hi-index | 0.00 |
A new technique of color face recognition is proposed. First processing stage consists of an optimum color conversion from the 3D RGB space into a 2D selected feature space using the old Karhunen-Loève transform (KLT). The resulted 2D color space is defined by the two color components (called C1 and C2), corresponding to the two largest eigenvalues of the RGB pixel covariance matrix. The second processing phase corresponds to Principal Component Analysis (PCA) for each color channel. Third stage corresponds to the feature fusion of the C1 and C2 PCA-components. Last processing stage is a multiple neural classifier consisting of a set of concurrent self-organizing modules. The proposed system is experimented for the Essex color face database containing 3520 color images of 151 subjects.