基于模糊模式和BP算法的手写数字识别  

Handwritten Digit Recognition Based on Fuzzy Pattern and BP Arithmetic

作　　者：刘荣华[1] 徐学洲[1] 

机构地区：[1]西安电子科技大学软件学院,陕西西安710071

出　　处：《电子科技》2006年第3期60-63,共4页Electronic Science and Technology

摘　　要：通过对多种手写数字识别算法的研究和分析,提出了一种基于模糊模式识别和BP神经网络对手写数字进行识别的新算法。首先应用BP神经网络对手写数字样本进行学习,然后再结合模糊模式识别的思想进行手写数字识别。实验证明:该算法与传统的手写数字识别算法相比,识别率明显提高。经过推广,该算法可应用于汉字识别、人脸识别等领域。By studying and analyzing many algorithms for handwritten digit recognition, this paper proposes a new algorithm for handwritten digit recognition based on fuzzy pattern recognition and BP neural network, First, the new algorithm applies the BP neural network to study the digital sample. Second, during the recognition of handwritten digit, the new algorithm applies the thought of fuzzy pattern recognition. The experiment proves that this algorithm greatly improves recognition. This algorithm can be applied to the recognition of Chinese characters, human face, etc through popularization.

关 键 词：模糊识别 模式识别 数字识别 神经网络 BP算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]