一种基于微结构特征的多文种文本无关笔迹鉴别方法

与字符识别一样, 计算机自动笔迹鉴别是一个涉及到不同文种的研究课题. 本文提出了一种基于网格窗口微结构特征的文本无关的笔迹鉴别方法, 能适用于各种不同文种的笔迹. 该方法对笔迹中局部细微结构的书写变化趋势进行描述, 并采用加权距离度量方法进行笔迹相似性度量. 利用该方法实现了文本无关的多文种笔迹检索系统, 并在实际汉字、英文、藏文和维吾尔文的笔迹库上进行了测试. 实验证明, 该方法是一种高效且适用性较广、限制性较少的笔迹鉴别方法.

A Microstructure Feature Based Text-independent Method of Writer Identification for Multilingual Handwritings

As the same as character recognition, computer automatic writer identification is a research subject involving different languages. This paper proposes a text-independent method of writer identification based on grid-window microstructure feature for different multilingual handwritings. The proposed method depicts the writing trend of local fine structures in handwritings and uses weighted distance metrics to measure the similarity between handwritings. Based on the proposed method, a text-independent handwriting retrieval system is implemented. The system is tested on the real handwriting databases of Chinese handwriting, English handwriting, Tibetan handwriting, and Uighur handwriting. The experimental results demonstrate that the proposed method is a general-purpose and weak-confined method of writer identification with high efficiency.