摘要
为了解决ART2神经网络的漂移问题,提出了一种改进的基于ART2神经网络的文字分类和识别方法。此方法能够自主学习,收敛速度快,识别率和识别速度都比BP神经网络高。实践证明,基于此设计的脱机手写体文字识别系统能对较规范的手写体文字进行识别,识别率达到85%。
An improved character recognition algorithm is improved, which mainly solves the problem of shift in ART2 neural network. The method achieves unsupervised learning, faster convergence velocity, higher recognition rate than BP neural network. The practice has proved that the offline handwritten character recognition system based on the improved algorithm can recognize normative handwritten characters. The recognition rate is up to 85%.
出处
《桂林电子科技大学学报》
2012年第3期237-239,共3页
Journal of Guilin University of Electronic Technology
基金
广西高等教育教学改革工程(2012JGA240)
