Face Recognition Using State Space Parameters and Artificial Neural Network Classifier

This paper presents a new approach to model face images using the state space feature parameters. We present a novel feature extraction method for the recognition of face images based on their grayscale images eliminating any step of pre-processing. Experiments are performed using the standard AT & T (formerly, ORL face database) face database containing 400 face images of 40 different individuals. The sate space map and state space point distribution graph drawn for 400 individuals' face image shows the credibility of the method. To show the nonlinear nature of the face images the fractal dimension is also computed from the sate space map of the each face image using the box count method. In the recognition stage we used artificial neural network classifier, and the proposed SSPD feature is found to be promising, and this is the first attempt of this kind in the field of face recognition.