Face verification by simple networks

In this paper we used an arrangement of simple neural networks to provide a new representation of patterns to verify frontal pictures. For this purpose, in the training phase, the client set is repeatedly divided into two specified sub-sets(super-classes) to train a set of binary experts (neural networks)using Error correcting output coding technique. A modified version of ECOC (multi-seed ECOC), based on using several representative (seeds) for each class, is suggested for verification, which exhibits better performance on the well known XM2VTS data set compared with previous reported results.

Keywords:

verification , multi-seed ECOC , combining classifiers