基于粗集方法的一种文字识别

METHOD OF WORD RECOGNITION ON ROUGH SETS

文字识别作为模式识别的一个重要方面，其研究获得广泛重视。本文基于粗集理论的观点，把文字识别问题看作是模式识别中简化知识表达与信息分类处理问题，提出一种新的文字识别方法，这不仅对模式识别中如何提取有用特征是一种新的探索，而且对简化信息处理也是一种新的方法。本文介绍了基于粗集理论的文字知识表达的方法，探讨了模式识别中的特征提取及最小决策算法，并给出一个文字识别的例子来说明这种有效的方法。

As an important discipline of pattern recognition, word recognition is extensively studied. From the view of rough sets,word recognition is considered to simplifications of knowledge representation and process of signal classifcation. We propose a new method of word recognition which is not only a new kind of investigation on extraction of useful characteristics of pattern recognized,but also a novel approach on reduction of space dimension of knowledge representation. In this paper,simplification and approximation of knowledge representation of words on rough sets are introduced,and a minimum decision algorithm of word recognition is investigated. Finally,a practical example of application on word recognition is given to illustrate the efficient method proposed here.