Off Line Handwritten Signature Verification Based on Feature Fusion

At present most of the research on offline handwritten signature is based on a single language and the problems of the sparse signature image, weak feature representation ability and low verification rate have not been well solved. In this paper, the off-line handwritten signature images of two different languages including Chinese and Kazakh are used as experimental data. the experimental results show that even a small amount of training data. The accuracy rate of this paper in multi-lingual off-line handwritten signature verification can still reach 96.74% compared with related work the verification effect of this method is higher.