Analyzing wavelets components to perform face recognition

Proceeding of: 6th International Work-Conference on Artificial and Natural Neural Networks, IWANN 2001 Granada, Spain, June 13–15, 2001Face recognition is a very difficult task in real environments. In those cases a good preprocessing of the images is needed to keep the images invariant to translations, scales, luminosity, shape, aspect, rotation, noise, etc ... Wavelet transformation have been probed to be a good preprocessing method for many task. However, not all the coefficients of a wavelet transform have the information needed for a classification method to be efficient. This work introduce a method to select the most appropriate coefficients for a wavelet transform to allow an unsupervised neural network to well classify a set of complex faces