On-line signature verification based on subband decomposition by DWT and adaptive signal processing

This paper presents an on-line signature verification method based on the Discrete Wavelet Transform (DWT) and adaptive signal processing. Time-varying pen-position signals of the on-line signature are decomposed into subband signals by using the DWT. Individual features are extracted as high-frequency signals in subbands. Verification is achieved in each subband by determining whether the adaptive weight converges to unity. The overall verification decision is made by averaging the verification results in lower subbands. Experimental results show that the verification rate was 95% even if the writer was not permitted to refer to his/her own signature or if a forger traced the genuine signature. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(6): 1–11, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20143