Arabic/Persian cursive signature recognition and verification using line segment distribution

摘要

This work proposes a fast method for line segment extraction based on chain code differentiation. It is applied to cursive signature recognition of Arabic/Persian. The evaluation method is introduced to obtain a quantitative value for the recognition rate. The comparative results show the existing differences among the methods in recognition, building time and searching time criteria. The two methods used for comparison are invariant moments and CBLSE method.