VALSE2019总结(5)-一些很有insight的poster宣讲

4.9 一些很有insight的poster宣讲

• Independently Recurrent Neural Networks (IndRNN, NIPS2018)

  

  

  

  

• FishNet ()

  

  

  

• HRNet: Deep High-Resolution Representation Learning for Human Pose Estimation (CVPR 2019, msra)

  

  • HRNet有点像 UNet的终极版本：多分辨率的巧妙结合(尽可能获得双高：高语义高分辨率)，motivation极其简单(希望双高，不希望从低分辨率上采样到高分辨)，但是效果极其好(图像分割中效果拔群)

  

• 王井东(MSRA)

  

  

  

  

  

  

  

• Text Guided person image synthesis (文字指导图片中人物的姿态变化)

  

  

• PCGAN

  

• Graphical GAN (NIPS 2018)

  

4.10 有趣的Poster

• GLCN (CVPR2019): semi-supervised learning with graph learning-convolutional networks

  

  

  

• Data Representation and Learning with graph diffusion-embedding networks (GCN的一种)

  

• multi-label image recognition with GCN

  

• two-stream adaptive GCN for skeleton-based action recognition

  

• An attention Enhanced Graph Convoulutional LSTM Network for sketelon-based Action recognition

  

• All-in-focus synthetic aperture imaging using image matting