Comparative Analysis of Global Feature Extraction Methods for Off-line Signature Recognition

Abstract

this paper, a method is proposed for feature extraction of offline signature recognition system. The proposed method is based on global features to identify forgeries and also median filter is introduces for noise reduction. The Proposed feature extraction method is compared with Discrete Radon Transform (DRT). Both the feature extraction method extracts one dimensional global features and the alignment between features is performed by Dynamic Time Warping (DTW). When being trained using 6 genuine signatures of each person and 250 forgeries taken from our database, the proposed method obtained an equal error rate (EER) of 8.40%. The false acceptance rate (FAR) for proposed method was also kept as low as 8.80%. Keywordsfilter, Global features, Discrete Radon transform, Dynamic time warping.