Signature Verification by using Radial Basis Function (SVM)
- February 6, 2018
- Posted by: RSIS
- Categories: Computer Science and Engineering, Engineering
International Journal of Research and Innovation in Applied Science (IJRIAS) | Volume I, Issue I, April 2016 | ISSN 2454-6194
Signature Verification by using Radial Basis Function (SVM)
Gurpreet Singh#, Gurbinder Singh*, Mandeep Singh#
Department Research Scholar, AIET, Faridkot
#Department of CSE, PTU, AIET, Faridkot
Abstract:- Computerization has definitely revolutionized the way banking is done these days. However, even today all banking transactions, especially, financial require our signatures to be authenticated. The identifiable side-effect of signatures is that they are vulnerable to forgery. In order to avoid misuse or manipulations on account of forged signatures the need for research in efficient automated solutions for signature recognition and verification has increased in recent years. A concrete system has to be developed which should not only be able to consider these factors but also detect various types of forgeries. Signature verification approaches using information technology can be categorized according to the acquisition of the data: On-line and Off-line. On-line data records the motion of the stylus while the signature is produced, and includes location, and possibly velocity, acceleration and pen pressure, as functions of time. Online systems could be used in real time applications like credit cards transaction or resource access. While off-Line signature verification systems take as input the 2-D image of a signature. Offline systems are useful in automatic verification of signatures found on bank checks and documents. Artificial Neural Network (ANN) which has been modeled on human brain has been successfully used classifier in numerous fields. The present study focuses on detection of these forgeries using Support Vector Machine with Radial Basis Function (SVM – RBF) kernel. For evaluating our system’s performance and to know the output unit response accuracy we have developed a classification/confusion matrix and kept our performance goal based on mean square error. It was found that while ANN base system had an overall accuracy of 90% the accuracy of SVM –RBF kernel was significantly higher at 95%.
Keywords: – Signature Verification, Radial Basis Function, neural network, SVM etc.
