Online Signature Verification: Automatic Feature Selection vs. FHEs Choice

Abstract

In this paper, the discriminative power of a set of features which seems to be relevant to signature analysis by Forensic Handwriting Experts (FHEs) is analysed and particularly compared to the discriminative power of automatically selected feature sets. This analysis could help FHEs to further understand the signatures and the writer behaviour. In addition, two information fusion schemes are proposed to combine the discriminative capability of the two types of features being considered. The coefficients in the wavelet decomposition of the different time functions associated with the signing process are used as features to model them. Two different signature styles are considered, namely,Western and Chinese, of one of the most recent publicly available Online Signature Databases. The experimental results are promising, especially for the features that seem to be relevant to FHEs, since the obtained verification error rates are comparable to the ones reported in the state-of-the-art over the same datasets. Further, the results also show that it is possible to combine both types of features to improve the verification performance.