Persian Signature Detection Based on PCA and MLP Neural Network

In this article , we investigate the performance of MLP (Multilayer Perceptron) network acting as a classifier in a signature recognition system. For this purpose, a data of 100 signature image of 10 people (10signatures for each person) is used for training and evaluation. The principal components of images have been used as inputs for the proposed network. The experimental results reveal that the MLP network exhibits appropriate performance with the recognition accuracy of 98.5%.

Keywords:

Multilayer Perceptron Neural Network , Principal Component Analysis , Signature Detection