Angle features extraction of handwritten signatures

The selection of the signatures' features is crucial for the success of any signature verification system. The features of signature can be divided into global features and local features. Local features represent a segment or limited region of the signature image. Although, they require more computations; they are much more accurate than global features. In this paper, feature extraction using angle features has been studied and implemented in system of handwritten signature verification in many methods which differ in the way of determining the width and the height of each part, and differ in determining the number of parts in dividing signature image. The efficiency of the system for each method has been tested on a local database. The local database of 880 signatures taken from 40 persons has been developed in this study.