A Performance Analysis of Face Recognition Based on Artificial Neural Network

Abstract

Face recognition is bearing in mind and recognizing a face of individual, like a visual pattern recognition that enables the system to verify or identify individual faces that is taken from imaging system. In this paper we tried to show that a face recognition that uses a tanh activation function would have a better or lower mean squared error instead of using the Logistic Sigmoid activation function. The network was trained several times on different ideal input and noisy images that are taken from imaging devices like camera.  In this case training a network on different sets of noisy images forced the network to learn how to deal with noise image by using a tanh activation function.