Filtros de Correlação e Características Invariantes à Escala para o Reconhecimento de Faces

The objective of this work is to combine, improve, and develop algorithms for face detection and recognition so as to create a software-based system which is able to detect and recognize faces, previously enrolled in a databasis, obtained from capture devices such as a webcam. In order to implemement such a system, two state-of-the-art algorithms were selected: CFA – Class-dependence Feature Analysis and SURF – Speed Up Robust Features, the last one being conceptually similar to SIFT – Scale Invariant Feature Transform. For CFA, it is proposed to use of DCT – Discrete Cosine Transform and KLT – Karhunen- Loève Transform as alternatives to the DFT, so as to reduce the time to compute the correlation filters and/or improve the ROC. For SURF, two configuration parameters are tested, having also as objective the determination of the best ROC.