Wavelet functions in face recognition systems

In this Master thesis studies the impact of wavelet transform on solving the common face recognition problems. Four data sets were created taking into account main face recognition problems: luminance changes, face cover, emotions and position changes. All four initial data sets were processed with wavelets („Harr”, „Coiflets”, „Symlets”, „Daubechies”) in order to detect edges and denoise images. Recognition result of LDA, PCA and ICA algorythms was computed for wavelet transformed data sets. Conclusions were also provided based on chosen evaluation criteria