PESC: Master Degree AbstractsAbstract of Thesis presented at COPPE/UFRJ as a partial fulfillment of the requirements for the degree of Master of Science (M.Sc.)
A System for Recognizing Human Face Using Image Processing and Bayesian Network
Paulo Sérgio de Souza Coelho
Antonio Alberto Fernandes de Oliveira
Department: Systems Engineering and Computer Science
This work describes an human face detectar system called Ldetec. This system was developed using Image Processing tools and a Bayesian Network. The face is modeled as a set of facial feature with defined position. Gaussian Filters are used to determine first facial feature candidates. These candidates was grouped in face candidates in accordance with the face modelo This grouping is based in self features of candidates and its position and gizes using statistical metrics to compare with previous treined data. Finally, the Bayesian Network is used to estimate probabilities of each face candidate be really a face. The results obtained show this system can detect face at different scales, orientations, and even with some partial occlusion.