基于LVQ神经网络算法的脱机手写数字识别研究

Off-line Handwritten Digit Recognition Based on LVQ Neural Network Research

针对BP神经网络进行脱机手写数字识别所存在的问题,提出用自组织竞争神经网络(LVQ)对脱机手写数字识别的方法.介绍了LVQ神经网络算法,并构建了LVQ神经网络的识别模型,用Matlab软件进行了仿真.将获得的仿真结果数据与BP神经网络的测试结果进行对比分析,发现LVQ神经网络对脱机手写数字的识别率明显高于BP神经网络,且收敛速度更快.该方法在脱机手写数字识别领域具有一定的可行性与指导性.

In view of the BP neural network for existing problems in the off-line handwritten digit recognition, This paper puts forward the method of off-line handwritten digital （LVQ） for off-line handwritten digit recognition. It introduces the LVQ neural network algorithm, and it constructs the LVQ neural network identification model, using Matlab software simulation. Comparison and analysis of the simulation results of the data obtained with the testing results of BP neural network, it found that the LVQ neural network recognition rate of off-line handwritten Numbers are significantly higher than the BP neural network and it has faster convergence speed. The method in the field of off-line handwritten digit recognition is of certain feasibility and guidance.