Optimized Signature Recognition System Based on Feature Extraction Methods and Neural Network Classifier

Vol. 3 No.2

Year: 2008

Issue: Oct-Dec

Title: Optimized Signature Recognition System Based on Feature Extraction Methods and Neural Network Classifier   

Author Name: Mohammed A. Abdala, Saja K. Jameel 

Synopsis:   

This paper presents new techniques for off-line signature recognition. The proposed systems are based on invariant moment, texture and global features. For each one of these features sets neural network classifiers which are based on MLP (Multilayer Perceptron), RBF (Radial Basis Function) and Elman neural network classifier are implemented. These nine signature recognition systems are tested then the optimized system that has the highest recognition rate is selected. The optimized system is evaluated with rotation and scaling effects.