CNN Visualizing and Understanding
Paper:Visualizing and Understanding Convolutional Networks
2014 ECCV 纽约大学 Matthew D. Zeiler, Rob Fergus
论文:Visualizing and Understanding Convolutional Networks(卷积神经网络的可视化理解)
论文下载:https://arxiv.org/pdf/1311.2901.pdf
论文翻译:https://blog.csdn.net/kklots/article/details/17136059
博客:http://kvfrans.com/visualizing-features-from-a-convolutional-neural-network/
文章主要技巧:使用可视化技巧——反卷积网络(Deconvnet)来直观看到CNN中间层的特征对应的图像区域,Deconvnet在(Zeiler et al., 2011)中被用作无监督学习,本文则用来进行可视化演示。一层反卷积网可以看做是一层卷积网的逆过程,它们拥有相同的卷积核和Pooling函数(准确说,是逆函数),因此反卷积网是将输出特征逆映射成输入信号。
论文:Adaptive Deconvolutional Networks for Mid and High Level Feature Learning
关于反卷积网络(Deconvnet):https://www.zhihu.com/question/43609045/answer/120266511