Anderson Rodrigo dos Santos,  Adilson Gonzaga. 

The use of biometric technique for automatic personal identification is one of the biggest challengesin the security field. The process is complex, because it is influenced by many factors related tothe form, position, illumination, rotation, translation, disguise and occlusion of face characteristics.This work presents a searching method to identify a face in a training database. We have proposedan algorithm for face recognition based on LDA subspace using a SOM neural net to memorizeeach class (face) in the stage of classification/ identification. The interaction between the number of eigenvectors in the PCA and LDA subspaces has been analyzed to establish the rate recognition.

http://www.lbd.dcc.ufmg.br/colecoes/wvc/2005/005.pdf

Processando consulta.