On-Line Signature Verification Utilizing 3-Axis Force Features Extraction Based on DWT

Abstract

Automatic on-line signature verification is an intriguing intellectual challenge with wide attention and practical applications. We examined the validation of the new dynamic features for on-line signature verification, which extracted from 3-axis forces of pen-tip to writing tablet with application of discrete wavelet transform (DWT). The force signals were decomposed into sub-band signals by DWT and individual features were extracted as locations of modulus maximum in detail and the values of the wavelet transform at the corresponding location. Verification was achieved by using dynamic time warping (DTW) and total decision was done by combining multiple thresholds. Experimental results indicate the effectiveness of the dynamic features for on-line signature verification described in this paper.