On the usage of i-vector representation for online handwritten signature verification

摘要

Handwritten signature verification is one of the common methods used by banks, intelligence agencies and high-profile institutions to validate the identity of an individual. It is done in two manners: offline and online. By technological advancement, online methods are considered more in literature and several methods have been proposed for this task. In this paper, a new method based on i-vector is proposed for online signature verification. This method was originally proposed for speaker verification and we adopt it for this area. In this method, a fixed length vector is extracted from each signature called i-vector and then this vector is used for template making. Scoring is done using cosine similarity. Experimental results on SigWiComp2013 database show the excellence of the proposed method compared to several other methods. In addition to the superior accuracy, this method is much faster than other methods.