在线签名鉴别的改进DTW算法

Improved DTW algorithm applying on on-line signature verification

论述一种改进的在线签名鉴别算法,在原有的应用DTW算法基础上,在所需鉴别签名的动态特征中加入角度特征,克服了规整中将两个签名中的同一笔画规整成不同的笔画的弊端,提高了鉴别的准确率.同时,提出的选取最优化样本的方法更好地增强了在线签名的实时性,并且通过反复试验得以验证.

This paper discusses an improving algorithm of online handwriting signature. Based on primary DTW arithmetic, it overcomes the defect which results in warping the same stroke to other strokes in two signatures. The paper adds the angle characters to the dynamic characters in order to improve the veracity. At the same time, also adopts the optimized sampling method choosing best sample to enhance the real time performance of online handwriting signature which is proven by repeated tests.