A principal component based probabilistic DBNN for face recognition

Principal component analysis (PCA) is a powerful statistical approach for extracting facial features for recognition. The eigenface method has been reported to provide significant recognition performance over various testing and evaluation procedures. We try to improve the PCA recognition performance by concatenating a probabilistic decision based neural networks (DBNN). Our experiments show that the hybrid PCA/NN systems can improve the recognition rate by about 8% better than the PCA systems, on our facial database, which contains large rotation face images as the testing sets.