CVPR2022论文速递（2022.4.5）！共28篇

整理：AI算法与图像处理

CVPR2022论文和代码整理：https://github.com/DWCTOD/CVPR2022-Papers-with-Code-Demo

分类 / Classification - 1 篇

Matching Feature Sets for Few-Shot Image Classification

• /Paper: http://arxiv.org/pdf/2204.00949
• /Code: None

目标检测 / Object Detection - 2 篇

R(Det)^2: Randomized Decision Routing for Object Detection

• /Paper: http://arxiv.org/pdf/2204.00794
• /Code: None

Homography Loss for Monocular 3D Object Detection

• /Paper: http://arxiv.org/pdf/2204.00754
• /Code: https://github.com/gujiaqivadin/HomographyLoss

目标跟踪/Object Tracking - 1 篇

Unsupervised Learning of Accurate Siamese Tracking

• /Paper: http://arxiv.org/pdf/2204.01475
• /Code: https://github.com/florinshum/ulast

语义分割/Segmentation - 3 篇

FIFO: Learning Fog-invariant Features for Foggy Scene Segmentation

• /Paper: http://arxiv.org/pdf/2204.01587
• /Code: None

WildNet: Learning Domain Generalized Semantic Segmentation from the Wild

• /Paper: http://arxiv.org/pdf/2204.01446
• /Code: https://github.com/suhyeonlee/wildnet

Semantic-Aware Domain Generalized Segmentation

• /Paper: http://arxiv.org/pdf/2204.00822
• /Code: https://github.com/leolyj/san-saw

Transformers - 2 篇

TransRAC: Encoding Multi-scale Temporal Correlation with Transformers for Repetitive Action Counting

• /Paper: http://arxiv.org/pdf/2204.01018
• /Code: https://github.com/sviprepetitioncounting/transrac

Consistency driven Sequential Transformers Attention Model for Partially Observable Scenes

• /Paper: http://arxiv.org/pdf/2204.00656
• /Code: None

多模态 / Multimodal - 1 篇

STCrowd: A Multimodal Dataset for Pedestrian Perception in Crowded Scenes

• /Paper: http://arxiv.org/pdf/2204.01026
• /Code: https://github.com/4dvlab/stcrowd

其他/Other - 18 篇

Joint Hand Motion and Interaction Hotspots Prediction from Egocentric Videos

• /Paper: http://arxiv.org/pdf/2204.01696
• /Code: None

LISA: Learning Implicit Shape and Appearance of Hands

• /Paper: http://arxiv.org/pdf/2204.01695
• /Code: None

Exemplar-bsaed Pattern Synthesis with Implicit Periodic Field Network

• /Paper: http://arxiv.org/pdf/2204.01671
• /Code: None

Correlation Verification for Image Retrieval

• /Paper: http://arxiv.org/pdf/2204.01458
• /Code: https://github.com/sungonce/cvnet

Degradation-agnostic Correspondence from Resolution-asymmetric Stereo

• /Paper: http://arxiv.org/pdf/2204.01429
• /Code: None

RayMVSNet: Learning Ray-based 1D Implicit Fields for Accurate Multi-View Stereo

• /Paper: http://arxiv.org/pdf/2204.01320
• /Code: None

Exploiting Temporal Relations on Radar Perception for Autonomous Driving

• /Paper: http://arxiv.org/pdf/2204.01184
• /Code: None

BNV-Fusion: Dense 3D Reconstruction using Bi-level Neural Volume Fusion

• /Paper: http://arxiv.org/pdf/2204.01139
• /Code: None

ES6D: A Computation Efficient and Symmetry-Aware 6D Pose Regression Framework

• /Paper: http://arxiv.org/pdf/2204.01080
• /Code: None

Style-Based Global Appearance Flow for Virtual Try-On

• /Paper: http://arxiv.org/pdf/2204.01046
• /Code: https://github.com/senhe/flow-style-vton

Neural Global Shutter: Learn to Restore Video from a Rolling Shutter Camera with Global Reset Feature

• /Paper: http://arxiv.org/pdf/2204.00974
• /Code: https://github.com/lightchaserx/neural-global-shutter

DST: Dynamic Substitute Training for Data-free Black-box Attack

• /Paper: http://arxiv.org/pdf/2204.00972
• /Code: None

AdaFace: Quality Adaptive Margin for Face Recognition

• /Paper: http://arxiv.org/pdf/2204.00964
• /Code: https://github.com/mk-minchul/adaface

Progressive Minimal Path Method with Embedded CNN

• /Paper: http://arxiv.org/pdf/2204.00944
• /Code: None

Online Convolutional Re-parameterization

• /Paper: http://arxiv.org/pdf/2204.00826
• /Code: None

What to look at and where: Semantic and Spatial Refined Transformer for detecting human-object interactions

• /Paper: http://arxiv.org/pdf/2204.00746
• /Code: None

Robust Neonatal Face Detection in Real-world Clinical Settings

• /Paper: http://arxiv.org/pdf/2204.00655
• /Code: None

SIMBAR: Single Image-Based Scene Relighting For Effective Data Augmentation For Automated Driving Vision Tasks

• /Paper: http://arxiv.org/pdf/2204.00644
• /Code: None