一种基于证据理论的手写体数字字符识别方法  

A Script Figure Character Recognition Method Based on the Evidence Theory

作　　者：王瑜[1] 

机构地区：[1]湖北大学数学与计算机科学学院,武汉430062

出　　处：《计算机工程与应用》2001年第5期59-61,共3页Computer Engineering and Applications

摘　　要：证据理论是一种适用于不确定决策问题和风险型决策问题的决策方法。它将各种从不同角度出发的意见综合起来，作用于识别框架上，得到问题的合理解。文章将其引入模式识别领域中，针对手写体数字字符识别问题，文章给出了用证据理论方法解决的具体模型和实例。这种方法对于研究和解决其它模式识别，决策问题具有较高的借鉴意义。： The evidence theory is a decision－making method used to solve uncertain and risk decision－making problems．It deduces a reasonable explain by intergating all opinions which interpret the problem from different viewpoints，and acting them on the recognization frame．This paper introduces the evidence theory into the area of pattern recognization．To solve the script figure character recognization problem，a substaintial model and an instance using the evidence theory are given in this paper．This method has some value in term with other pattern recognization and decision－making problems.

关 键 词：证据理论 手写体数字字符识别 模式识别 神经网络 信度函数 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]