这篇文章是CVPR2017的oral，非常厉害。文章提出的DenseNet（Dense Convolutional Network）主要还是和ResNet及Inception网络做对比，思想上有借鉴，但却是全新的结构，网络结构并不复杂，却非常有效！众所周知，最近一两年卷积神经网络提高效果的方向，要么深（比如ResNet，解决了网络深时候的梯度消失问题）要么宽（比如GoogleNet的Inception），而作者则是从feature入手，通过对feature的极致利用达到更好的效果和更少的参数。(This article is the oral of CVPR2017, very powerful. This paper DenseNet (Dense Convolutional Network) mainly and ResNet and Inception Network do contrast, reference ideas, but he is a new structure, Network structure is not complicated, but very effective! As is known to all, in the last year or two, the direction of improving the effect of convolutional neural network is either deep (such as ResNet, which solves the gradient disappearance problem in deep network) or wide (such as GoogleNet's Inception), while the author starts from feature and achieves better effect and fewer parameters through the extreme utilization of feature)