【源码】开源卷积神经网络工具箱ver 0.83

本工具箱提供了卷积神经网络的MATLAB实现代码。

This project provides matlab class forimplementation of convolutional neural networks.

该类神经网络由YannLeCun开发，并已成功地用于许多实际应用，如手写数字识别、人脸检测、机器人导航等。

This networks was developed by Yann LeCunand have sucessfully used in many practical applications, such as handwrittendigits recognition, face detection, robot navigation and others (see referencesfor more info).

由于卷积神经网络的一些结构特征，如权值共享等，如果不对其源代码进行修改，就不可能使用Matlab自带的神经网络工具箱来实现它。（MATLAB自带的神经网络工具箱是无法访问源码的！！！）

Because of some architectural features ofconvolutional networks, such as weight sharing it is imposible to implement itusing Matlab Neural Network Toolbox without it’s source modifications.

这也是我们单独设计这个独立于神经网络工具箱的原因。

That’s why this class works almostindependently from NN toolbox (coming soon full independence).

这个版本包含了使用CNN进行手写数字样本识别的案例。

This release includes sample of handwrittendigits recognition using CNN.

如果你想运行这个例子，请输入cnet_tool命令。

If you just want to try it run cnet_tool.

然后，你会看到一个简单的GUI界面。

You’ll see a simple GUI.

该界面从cnet.mat中加载预训练卷积神经网络，从MNIST数据库下载或在绘图区域进行数字图像识别。

It loads pretrained convolutional neuralnet from cnet.mat and recognizes image of digit either painted in painting areaor downloaded from MNIST database.

MATLAB源码下载地址：

http://page5.dfpan.com/fs/5lcjd2211291969bf32/

更多精彩文章请关注微信号：