一个高速准确的手写数字识别系统

对于手体写数字的识别提出了全新思路，阐述了整个识别步骤中的关键算法。系统可以自动解析包含二值化手写体数字的BMP文件，在经过裁剪、细化、映射归一、获取数字结构特征、匹配等一系列处理之后，可以最大限度地减少因为不同手写风格而带来的干扰，从而快速准确地识别出每个数字。基于所描述的算法而实现的程序，有着很好的识别性能，应用前景十分广泛。

A High Performance System for the Recognition of Handwritten Figures

This paper proposes a new method for the recognition of handwritten numerals, which details the key algorithms of the whole process. The system can automatically analyses bitmap files which contain the binarized handwritten digits. By processing each figure such as cutting, thinning, mapping and normalizing,obtaining the construction feature of the figure, and matching, we can reduce the confusion due to the different handwritten styles, then complete the recognition of the handwritten numerals quickly and accurately. The paper describes a new and creative idea and method for the handwritten digit recognition. The program basing on this algorithm runs quickly and the recognition result is satisfied. It is good for applying in many fields.