Adaptive Feature Extraction Algorithm using Mixed Transforms for Facial Recognition
Genevieve Sapijaszko, Taif Alobaidi, Wasfy B. Mikhael · 2018
An essential first step in facial recognition is the extraction of unique and reliable features that can identify faces from images. Feature extraction algorithms have evolved in recent years, with Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT) in particular showing good results. This paper utilizes the DWT and the DCT in sequence and iteratively to best find the features that represent the facial image. The proposed facial recognition system will use the ORL, Yale and the Ferret-Fc databases to compare the proposed system to different published results as well as a simplified version of the proposed system. Each of the feature extraction matrix for each image will be compared using the L1 norm classifier, and the best recognition rate will be determined.