On-line signature verification based on support vector data description and genetic algorithm

With the development of pen-based mobile device, on-line signature verification is gradually becoming a promising kind of biometrics. A method for the verification of on-line handwritten signatures using both support vector data description (SVDD) and genetic algorithm (GA) is described. A 27-parameter feature set including the shape and dynamic features was extracted from the on-line signatures data. The genuine signatures of each subject were treated as target data to train the SVDD classifier. As a kernel based one-class classifier, SVDD could accurately describe the feature distribution of the genuine signatures and detect the forgeries. To improve the performance of the verification, the feature subset selection and the parameters of classifier were jointly optimized by GA. Signature data form the SVC2004 database were used to carry out verification experiments. The proposed method has 4.93% average Equal Error Rate (EER) for skill forgery database.