应用图论和基元方向信息的手写数字识别

Handwritten Numeral Recognition Using Graph Theory and Direction Information

提出一种应用图论的原理和基元方向信息来识别手写数字的方法,根据图论的原理对数字的图像进行粗分类,抽取基元,由方向信息进行细分类,结果表明该方法有比较好的识别结果。

An approach is presented which applies the principle of graph theory and direction information to handwritten numeral recognition. It classifies crudely using graph theory. Primitives are extracted by direction information. The result shows that the approach can get a higher recognition rate.