Design and development of the digital implementation of a multilevel feed forward neural network architecture for face recognition based on statistical features representing Eigenfaces is presented. The architecture is divided into three parts viz. feature extractor, classifier and identifier. The Eigenface extractor architecture is developed based on an efficient design strategy in which all the M weight values corresponding to the Eigenfaces are generated simultaneously from M images representing the Eigen vectors and the test input image. The multilayer neural network classifier is trained using error back propagation algorithm. A novel multilevel digital architecture is developed for the implementation of the multilayer feed forward neural network for categorization of the input vectors into specific output classes. At the output of the back propagation network, a maximization network is used for the final classification of the multilevel outputs from the neural network. The data interpretation concepts adopted in the system design led to an efficient design methodology, which eliminated the necessity of complex computations needed for the implementation of multilayer perceptron using sigmoid function.

References

[1]    Gaeda R., Cerda J., Ballester F., and Mocholi A., Artificial neural network implementation on a single FPGA of a pipelined on-line back-propagation, IEEE 13th International Symposium on System Synthesis (ISSS'00), (2000) 225-230.

[2]    Lawrence S., Lee Giles C., Tsoi A.C., and Back A.D., Face recognition: A convolution neural network approach, IEEE Transactions on Neural Network, 8 (1997) 98-113.

[3]    Turk M.A., and Pentland A.P., Face recognition using eigenfaces, IEEE Conference on Computer Vision and Pattern Recognition, (1991) 586-591.

[4]    Belhumeur P.N., Hespanha J.P., and Kriegman D.J., Eigenfaces vs. Fisherfaces: Recognition using class specific linear projection, IEEE Transactions on Pattern Analysis and Machine Intelligence, 19 (7) (1997) 711-720.

[5]    Asari K. V., Srikanthan T., Kumar S., and Radhakrishnan D., A Pipelined Architecture for Image Segmentation by Adaptive Progressive Thresholding, Journal of Microprocessors and Microsystems, 23 (1999) 493-499.

[6]    Asari K. V., Srikanthan T., Kumar S., and Radhakrishnan D., A Pipelined Architecture for Image Segmentation by Adaptive Progressive Thresholding, Journal of Microprocessors and Microsystems, 23 (1999) 493-499.

[7]    Asari K.V., Non-linear spatial warping of endoscopic images: an architectural perspective for real-time applications, Journal of Microprocessors and Microsystems, 26 (2002) 161-171.

[8]    Optican, L.M. & Richmond B.A., Temporal encoding of two-dimensional patterns by single units in primate inferior temporal cortex. III - Information theoretic analysis, Journal of Neurophysiology, 57 (1984) 1143-1148.

[9]    Maher M.A.C., Deweerth S.P., Mahowald M.A. & Mead, C. A., Implementing neural architectures using analog VLSI circuits, IEEE Trans Circuits Systems, 36 (1989) 643﷓652.

[10]    Bayraktaroglu I., Ogrenci A.S., Dundar G., Balkir S., and Alpaydin E., ANNSyS: An analog neural network synthesis system, Neural Networks, 2 (1999) 325-338.

[11]    Botros N. M., and Abdul-Aziz M., Hardware implementation of an artificial neural network, Proc. ICNN, 3 (1993) 1252-1257.

[12]    Vijayan Asari, K., and Eswaran C., Systolic array implementation of artificial neural networks, Journal of Microprocessors and Microsystems, 18 (8) (1994) 481-488.

[13]    Erten G., Rodney M., and Goodman, A digital neural network architecture using random pulse trains, Proc. International Joint Conference on Neural Network, 1(1992) I-190- I-195.

[14]    Eldredge J.G., and Hutchings B.L., RRANN: A hardware implementation of the back propagation algorithm using re-configurable FPGAs, Proceedings of the IEEE World Conference on Computational Intelligence, (1994) 2097-2102.

[15]    Cimpu V.F., Hardware FPGA implementation of a neural network, Proc. Int. Conf. Technical Informatics, 2 (1996) 57-68.