CVPR2019

CVPR论文总结

Best Paper： 

A Theory of Fermat Paths for Non-Line-of-Sight Shape Reconstruction

非视线形状重建的费马路径理论  搞机器人的那一套不是很懂，但感觉到机器人很火，后续发展方向

卡内基梅隆 辛书冕

Best Student：

Reinforced Cross-Modal Matching and Self-Supervised Imitation Learning for Vision-Language Navigation

面向语言视觉导航（VLN）的强化交叉模型匹配和半监督模仿学习，做视觉语言导航的，也不是很懂，技术交叉有点多，也可以算是机器人方向

杜克大学 王鑫

经典论文奖 李飞飞～  ImageNet 大名鼎鼎 名副其实

 

最佳论文荣誉提名

英伟达的 A Style-Based Generator Architecture for Generative Adversarial Networks

谷歌的 Learning the Depths of Moving People by Watching Frozen People

人脸篇：

1. ArcFace oral
   提出ArcFace loss 从角空间进行约束，比较了同类型的角空间loss，证明自己的优越性
2. Deep Tree Learning for Zero-shot Face Anti-Spoofing    oral
3. APDrawingGAN: Generating Artistic Portrait Drawings from Face Photos with Hierarchical GANs
4. High-Quality Face Capture Using Anatomical Muscles
5. AdaScale: Adaptively Scaling Cosine Logits for Effectively Learning Deep Face Representations
6. Combining 3D Morphable Models: A Largescale Face-and-Head Model

      7. MVF-Net: Multi-View 3D Face Morphable Model Regression

      8. Towards High-Fidelity Nonlinear 3D Face Morphable Model

      9. GANFIT: Generative Adversarial Network Fitting for High Fidelity 3D Face Reconstruction

 

      10. AdaptiveFace: Adaptive Margin and Sampling for Face Recognition 

           通过设置可以学习的Margin来解决数据样本不均导致不同类别的类内方差异所带来的问题

 

 

 

 

GAN篇：

    1. Geometry-Consistent Generative Adversarial Networks for One-Sided Unsupervised Domain Mapping oral

    2.  Label-Noise Robust Generative Adversarial Networks oral

    3. Sphere Generative Adversarial Network Based on Geometric Moment Matching oral

    4.  A Style-Based Generator Architecture for Generative Adversarial Networks oral

    5. Constrained Generative Adversarial Networks for Interactive Image Generation oral

 

1.Unsupervised Face Normalization with Extreme Pose and Expression in the Wild.

 UFN 目的为了得到矫正后的人脸，包括pose，expression，lighting的矫正。矫正后的人脸再去做分类性能得到明显提升。

 

 

NAS篇：

1. 1. MnasNet: Platform-Aware Neural Architecture Search for Mobile

表情篇：

A Compact Embedding for Facial Expression Similarity

Facial Emotion Distribution Learning by Exploiting Low-Rank Label Correlations Locally