基于知识的无约束手写数字识别系统

A Knowledge-Based Recognition System of Unconstrained Handwritten Numerals

研究一种基于知识的无约束手写数字识别系统．系统运用专家系统的知识表示、知识推理等方法，通过未知样本与知识库模糊匹配，得到识别结果．系统对随意书写的５０００个样本进行识别。

A knowledge-based recognition system of unconstrained handwritten numerals is proposed. The theory of expert system,such as knowledge expression,fuzzy inference is used in this system. The unknown sample is recognized by fuzzy matching with the rules in the knowledge base. 5 000 samples are tested in the system and an encouraging result is obtained.