摘要
随着深度学习概念的提出,深层神经网络成为机器学习的一个研究方向。在众多的神经网络模型中,卷积神经网络由于具有权值共享、局部感受野、降维等特点,取得了较好的应用成果。卷积神经网络结构包括输入层、卷积层、下采样层、全连接层、输出层,当前层的输出经过激活函数作用后成为下一层的输入。文章对卷积网络的数学模型进行详细推理,采用Python编码实现了深层卷积网络,分别在MNIST和CIFAR-10数据集上进行了测试。实验结果表明,无论是灰度图像还是彩色图像,卷积神经网络都具有较好的识别效果。
With the introduction of deep learning, deep neural networks have become a research direction of machine learning. In many neural network models, convolutional neural networks (CNNs) have achieved good application results because of its weight sharing, local receptive feld, and dimension reduction. CNNs consist of the input layer, the convolution layer, the down-sampling layer, the fully connected layer and the output layer. The output of the current layer becomes the input of the next layer after the activation function. In this paper, this paper describes the mathematical model of CNNs in detail and code CNNs by Python. It is tested on MNIST and CIFAR-10 datasets respectively. The experimental results show that CNNs have a good recognition effect on the gray and color images.
作者
李书清
Li Shuqing(Hainan College of Software Technology,Qionghai 571400,China)
出处
《无线互联科技》
2018年第19期41-43,共3页
Wireless Internet Technology
基金
海南省自然科学基金项目
项目名称:深度学习算法研究-基于双层卷积神经网络的泛函分析法
项目编号:20161009
海南省高等学校科学研究项目
项目名称:基于6LoWPAN的网络安全机制研究与实现
项目编号:Hnky2015-79