本文简单介绍了人工神经元网络的背景知识,提出了一种利用传统BP(Back-Propagation误差逆传播)网络识别印刷字符的方法,用C语言对其进行了实现.在进行了大量实验之后,实验结果表明该字符识别器具有较好的有效性和正确性,能够在合理的误差范围内以较好的效率成功识别字符.
This paper introduced the background knowledge of Artificial Neural Network and a way to use BP network to recognize printed characters was proposed, which was implemented with C language. After a lot of experiments, the experimental results show that the character recognizer has good validity and correctness. Besides, the printed characters can be recognized successful within the reasonable range of error.