Model Struct

　　Deep Residual Learning for Image Recognition(ResNet)

　　Residual Networks Behave Like Ensembles of Relatively Shallow Networks(ResNet的解释)

GAN

　　A Style-Based Generator Architecture for Generative Adversarial Networks(优化了G网络)

Style Transfer

　　Attention-aware Multi-stroke Style Transfer(训练时不需要style图)

Image Restoration

　　Dual Residual Networks Leveraging the Potential of Paired Operations for Image Restoration

Low-Level Vision

　　Learning Dual Convolutional Neural Networks for Low-Level Vision(S子网络恢复结构，D网络恢复细节)

Semi-Supervised

　　Semi-supervised Transfer Learning for Image Rain Removal(先有监督，再无监督+风格转换)

Tricks

　　Squeeze-and-Excitation Networks(SENet)