基于神经网络的手写字符识别优化研究

Research on the Optimization of Handwritten Character Recognition Based on Neural

随着人工智能和大数据的兴起,神经网络的应用越发广泛,使得机械识别成为一个热点研究问题。传统的手写字体识别是以BP神经网络为基础,在大样本的训练数据下对手写字体的进行智能识别。本文提出一种小样本抗干扰的识别方案,使得训练样本在数据较少且存在一定噪音干扰的情况下也能进行高精度、有效识别。首先通过对手写字符的预处理等步骤实现图像压缩,并构建BP神经网络。其次,通过实验数据对网络的拓扑结构进行分析,并结合遗传算法对BP神经网络的参数进行优化。最终,获得了合适的神经网络参数,使得识别网络具有收敛快、稳定性强的优点。

With the rise of artificial intelligence and big data, the application of Neural Networks has become more widespread and Mechanical Identification has become a hot topic. Traditional handwritten character recognition is based on the BP Neural Networks with a lot of data training to achieve convergence. This paper proposes an anti-jamming identification Neural Networks, which can achieve convergence by smaller data training. Firstly, the handwritten character images were compressed by pre-processing, and the BP Neural Network was construct. Secondly, the network topology was analyzed by using the data experiment, and parameters of the BP Neural Network were optimized by applying the genetic algorithms. Finally, we obtained the network parameters in which the neural network has the advantages of fast convergence and strong stability.