基于概率神经网络的手写体数字识别

Handwritten Digit Recognition Based on Probabilistic Neural Network

随着手写体数字识别技术的发展以及概率神经网络的应用,基于概率神经网络的手写体数字识别技术,即PNN技术,是手写体数字识别领域才刚刚开始的一个研究方向。本文把概率神经网络技术应用在数字识别系统中,在特征提取技术的基础上,设计了特征提取算法,通过手写体数字识别流程,构造了概率神经网络的分类器。最后在数据输入、特征提取、模型训练、测试等几个部分,实现了手写体数字识别,获得了令人满意的正确度。

With the development of digit recognition technology and neural network, handwritten digit recognition technology based on probabilistic neural network, namely, PNN technology, is just an inchoate research direction in the field of handwritten digit recognition. In this paper, the probabilistic neural network technology is applied in the digital identification system. Based on the technology of feature extraction, it designs the feature extraction algorithm. It constructs the probabilistic neural network classifier through handwrittcn numeral recognition process. Finally, it achieves the handwritten nt, meral recognition in the parts of data input, feature extraction, model training, testing and others, and it obtains a satisfactory degree of accuracy.