Backpropagation Neural Networks for Facial Recognition

Abstract

A survey was conducted testing the aptitude of neural networks to recognize human faces. The motivation for this a study is the possibility of designing automatic security systems. Pictures of 511 subjects were collected. The pictures captured both profiles and many natural expressions of the subject. Some of the subjects were wearing glasses, sunglasses or hats in some of the pictures. The images were compressed by a magnitude of 100, and converted into image vectors of 1400 pixels. The image vectors were fed into a back propagation neural network with one hidden layer and one output node. The networks were trained to recognize one target person, and to reject all other persons. Neural networks for 37 target subjects were trained using eight different training sets that consisted of 7000 or more unseen pictures. Our results indicate that a false acceptance rate of less than 1% can be obtained, and correct acceptance rate of 98% can be obtained when certain restrictions are followed. A review of earlier work done on face recognition is given, and a brief summary of the theory and methods of neural networks is included.