机构地区: 华南理工大学
出 处: 《工程图学学报》 1998年第4期75-80,共6页
摘 要: 针对手写体汉字识别的特点,本文提出了一种有监督竞争学习神经网络分类识别模型。此模型算法简单,结构简练。结合本人提出的手写体汉字的子块及其相关模糊特征提取方法,对国标28区手写体汉字进行了识别实验,取得了满意的效果. According to the characteristic of Handwritten Chinese Characters Recognition(HCCR), a supervisory competitive learning neural networks with simple structure, simple algorithm has been presented in this paper. Combining the block and its related funy feature extraction approach we presented before, a recognition experiment in the region 28th of the GB2312-80 had been taken, the experimental results had met our satisfaction.
领 域: [自动化与计算机技术] [自动化与计算机技术]