Simply Partitioned DP Matching and Threshold Equalizing in DWT Domain On-line Signature Verification

Abstract

In this paper, we introduce simply partitioned DP (dynamic programming) matching and threshold equalizing into the verification process in DWT (discrete wavelet transform) domain on-line signature verification in order to improve the performance. The simply partitioned DP matching divides both data series (verification and template data) into several partitions and calculate sub DP distance every partition. Even if mismatched pairs are caused in each partition, they are initialized at the beginning of the next partition and matching errors could be reduced. The threshold equalizing suppresses the variation range of optimal thresholds for all users (signatures), so that it prevents the verification performance from degrading by common use of single threshold for all users. In particular, we propose two equalizing methods in which the relation between the number of sampled data and optimal thresholds in signatures are approximated and adjusted by linear and nonlinear functions. In verification experiments using the signature database: SVC2004, it is confirmed that the proposed methods are effective for improving the performance.