摘要
特征抽取是图形识别的关键.采用模板匹配方式搜寻特征点,抽取特征信息,进而得到可识别的图形信息,并提出了一种基于神经网络的电气元件符号识别算法.实验结果表明,这种思路是可行及有效的.
Feature extraction is a key to figure recognition. With feature positions sought with a template matching method and feature message extracted, recognizable figure message is received. A algorithm of handdrawn electronic component symbol' recognition based on neural network is put forward. The experimental result indicates that the idea is actable.
出处
《江南大学学报（自然科学版）》
CAS
2003年第1期53-55,共3页
Joural of Jiangnan University (Natural Science Edition)
关键词
特征抽取
模板匹配
神经网络
算法
feature extraction
template matching
neural Network
Algorithm
