摘要
神经网络是当今人工智能的一个重要方面．本文对人工神经网络用于图形识别，尤其是在干扰环境下的识别能力进行了研究．通过对一组图形符号的训练识别，讨论了外加噪声与网络内部噪声对识别功能的影响，同时提出了一个ＢＰ网络训练的改进算法．
Nowadays neural network is an important aspect in artifical intelligence. This paper mainly discusses the recognition performance of the neural network, especially in noise background. A set of symbols were trained and classified, the influence to the recognition performace of BP network in adding external and internal noises was explored, also, an improved BP algorithm ware proposed.
出处
《微机发展》
1996年第2期28-31,共4页
Microcomputer Development
