Kernel Based Artificial Neural Network Technique to Enhance the Performance and Accuracy of On-Line Signature Recognition

Abstract

Signature recognition is a standout amongst the most essential biometrics verification strategies, is an indispensable piece of current business exercises, and is considered a noninvasive and non-undermining process. For online signature recognition, many techniques had been presented before. Nevertheless, accuracy of the recognition system is further to be improved and also equal error rate is further to be reduced. To solve these problems, a novel classification technique has to be proposed. In this paper, Kernel Based Artificial Neural Network (KANN) is presented for online signature recognition. For experimental analysis, two datasets are utilized that are ICDAR Deutsche and ACT college dataset. The proposed K-ANN classification method gives lower performance in terms of accuracy value with 66% TPR, 73% FPR in ACT and 50% TPR, 56% FPR in ICDAR datasets respectively.