关键词: 手写签名认证, 小波包分析, 特征提取, RBF神经网络

Abstract: It is difficult to extract the effective features of on-line signature verification, so a new method of dynamic feature extraction is proposed by decomposition and reconstruction of wavelet packet in this paper. The feature vectors that reflect the energy change of different frequency ranges are constructed by the method and the feature vectors can reflect the dynamic feature of signature effectively. Moreover, RBF neural network is used to recognize signature by the feature vectors constructed above. Experiment show that recognition is correct to 96.75% and equal error rate is 3.34%.The performance is satisfactory.

Key words: handwritten signature verification, wavelet packet analysis, feature extraction, RBF neural network