A Real-Time Signature Verification Technology Using Clustering and Statistical Analysis

Abstract

The aim of this paper is to present the detailed description of the working and application of ‘real-time signature verification’ using clustering and statistical analysis. The entitled system works faster and efficiently involving two processes, i.e., ‘preprocessing’ and ‘feature matching’ such that ‘noise reduction,’ ‘image normalization,’ and ‘skeletonization’ are carried out in preprocessing after which the input image is stored in the database. When the sample inputs new signature, ‘feature matching’ takes place, comprising brute force and sift algorithm, and matches the signature according to the cluster values present in the stored sample signature. The approach is fast efficient and authentic, which can be implemented in certain core areas. We have proposed integrated signature verification system, in which the brute force and sift algorithm are implemented in order to perform the matching. An integrated verification system not only provides a way to compare and match an online signature against an online signature but also improves the system performance in those cases where both static and dynamic features are available.