基于LeNet-5模型的手写数字识别优化方法  

Handwritten Digital Recognition Optimization Method Based on LeNet-5 Model

作　　者：汪雅琴 夏春蕾[1] 戴曙光[1] WANG Yaqin;XIA Chunlei;DAI Shuguang(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093)

机构地区：[1]上海理工大学光电信息与计算机工程学院

出　　处：《计算机与数字工程》2019年第12期3177-3181,共5页Computer & Digital Engineering

摘　　要：作为深度前馈人工神经网络的一种,卷积神经网络在图像识别领域得到了成功应用。其中,最经典的卷积神经网络模型就是LeNet-5模型。在MNIST字符库上运用该模型,通过优化卷积层的样本训练方式,即将原来以每批固定输入样本数量、固定迭代次数的训练方式,优化为以每批不同输入样本数量、不同迭代次数的混合训练样本方式。优化后的训练方式能够减少预处理工作量,加快识别速度。实验结果表明:在保证样本训练时间相等的前提下,优化后的混合样本输入方式可以得到更高的识别率。As a kind of depth feedforward artificial neural network,convolutional neural network has been successfully applied in the field of image recognition.Among them,the LeNet-5 model is the most classic convolutional neural network model.This model is used on the MNIST character library and the sample training method of convolution layer is optimized.That is to say,the training method that uses the number of fixed input samples per batch and the number of fixed iterations is optimized to be a mixed training sample mode with different numbers of input samples per batch and different iterations.The optimized training method can reduce the pre-processing workload and speed up the recognition speed.The experimental results show that the optimized mixed sample input method can get a higher recognition rate under the premise of equal sample training time.

关 键 词：图像识别 卷积神经网络 LeNet-5模型 MNIST字符库 手写数字识别 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]