Biometric based subregion analysis for handwritten signature

Subregions of signature area lead to the off-line signature analysis. It is observed in certain subregions of signature, the variation persists in different signing instances. By manual verification system, an acceptable level of variation can be noticed. The incipient applications for authenticatio...

Full description

Saved in:
Bibliographic Details
Published in2017 Innovations in Power and Advanced Computing Technologies (i-PACT) pp. 1 - 4
Main Author Abhinand, P.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.04.2017
Subjects
Authentication
Contour tracing
Feature extraction
Forgery
Gaussian-Hermite moment
Mathematical model
Quadtree decomposition
Smoothing methods
Splines (mathematics)
Zero-plot of inverse slope
Online AccessGet full text

Cover

More Information
Summary:Subregions of signature area lead to the off-line signature analysis. It is observed in certain subregions of signature, the variation persists in different signing instances. By manual verification system, an acceptable level of variation can be noticed. The incipient applications for authentication require simple robust methods to provide ideal substantiation. There exists natural variation in definite fragments of the signature provided by genuine subject at different intervals. Ignoring natural variation will dissipate the processing time in imposter samples. The design of verification system should consider all categories of forgeries. The selected list of quadtree components regulates the subregions. The proposed research work accomplish off-line signature verification with the classification techniques based on zero-plot of the inverse slope, contour tracing using intersection points and Gaussian-Hermite moment. The best results were achieved for method based on zero-plot of the inverse slope with false rejection rate (FRR) 4% and false acceptance rate (FAR) 5%.
DOI:10.1109/IPACT.2017.8244954