CVPR 2022 论文和开源项目合集

向AI转型的程序员都关注了这个号????????????

​机器学习AI算法工程   公众号：datayx​

【CVPR 2022 论文开源目录】

• Backbone

• CLIP

• GAN

• NAS

• NeRF

• Visual Transformer

• 视觉和语言(Vision-Language)

• 自监督学习(Self-supervised Learning)

• 数据增强(Data Augmentation)

• 目标检测(Object Detection)

• 目标跟踪(Visual Tracking)

• 语义分割(Semantic Segmentation)

• 实例分割(Instance Segmentation)

• 小样本分割(Few-Shot Segmentation)

• 视频理解(Video Understanding)

• 图像编辑(Image Editing)

• Low-level Vision

• 超分辨率(Super-Resolution)

• 3D点云(3D Point Cloud)

• 3D目标检测(3D Object Detection)

• 3D语义分割(3D Semantic Segmentation)

• 3D目标跟踪(3D Object Tracking)

• 3D人体姿态估计(3D Human Pose Estimation)

• 3D语义场景补全(3D Semantic Scene Completion)

• 3D重建(3D Reconstruction)

• 伪装物体检测(Camouflaged Object Detection)

• 深度估计(Depth Estimation)

• 立体匹配(Stereo Matching)

• 车道线检测(Lane Detection)

• 图像修复(Image Inpainting)

• 人群计数(Crowd Counting)

• 医学图像(Medical Image)

• 场景图生成(Scene Graph Generation)

• 弱监督物体检测(Weakly Supervised Object Localization)

• 高光谱图像重建(Hyperspectral Image Reconstruction)

• 水印(Watermarking)

• 数据集(Datasets)

• 新任务(New Tasks)

• 其他(Others)

Backbone

A ConvNet for the 2020s

• Paper: https://arxiv.org/abs/2201.03545

• Code: https://github.com/facebookresearch/ConvNeXt

• 中文解读：https://mp.weixin.qq.com/s/Xg5wPYExnvTqRo6s5-2cAw

Scaling Up Your Kernels to 31x31: Revisiting Large Kernel Design in CNNs

• Paper: https://arxiv.org/abs/2203.06717

• Code: https://github.com/megvii-research/RepLKNet

• Code2: https://github.com/DingXiaoH/RepLKNet-pytorch

• 中文解读：https://mp.weixin.qq.com/s/_qXyIQut-JRW6VvsjaQlFg

MPViT : Multi-Path Vision Transformer for Dense Prediction

• Paper: https://arxiv.org/abs/2112.11010

• Code: https://github.com/youngwanLEE/MPViT

• 中文解读: https://mp.weixin.qq.com/s/Q9-crEOz5IYzZaNoq8oXfg

CLIP

HairCLIP: Design Your Hair by Text and Reference Image

• Paper: https://arxiv.org/abs/2112.05142

• Code: https://github.com/wty-ustc/HairCLIP

PointCLIP: Point Cloud Understanding by CLIP

• Paper: https://arxiv.org/abs/2112.02413

• Code: https://github.com/ZrrSkywalker/PointCLIP

Blended Diffusion for Text-driven Editing of Natural Images

• Paper: https://arxiv.org/abs/2111.14818

• Code: https://github.com/omriav/blended-diffusion

GAN

SemanticStyleGAN: Learning Compositional Generative Priors for Controllable Image Synthesis and Editing

• Homepage: https://semanticstylegan.github.io/

• Paper: https://arxiv.org/abs/2112.02236

• Demo: https://semanticstylegan.github.io/videos/demo.mp4

Style Transformer for Image Inversion and Editing

• Paper: https://arxiv.org/abs/2203.07932

• Code: https://github.com/sapphire497/style-transformer

NAS

β-DARTS: Beta-Decay Regularization for Differentiable Architecture Search

• Paper: https://arxiv.org/abs/2203.01665

• Code: https://github.com/Sunshine-Ye/Beta-DARTS

ISNAS-DIP: Image-Specific Neural Architecture Search for Deep Image Prior

• Paper: https://arxiv.org/abs/2111.15362

• Code: None

NeRF

Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields

• Homepage: https://jonbarron.info/mipnerf360/

• Paper: https://arxiv.org/abs/2111.12077

• Demo: https://youtu.be/YStDS2-Ln1s

Point-NeRF: Point-based Neural Radiance Fields

• Homepage: https://xharlie.github.io/projects/project_sites/pointnerf/

• Paper: https://arxiv.org/abs/2201.08845

• Code: https://github.com/Xharlie/point-nerf

NeRF in the Dark: High Dynamic Range View Synthesis from Noisy Raw Images

• Paper: https://arxiv.org/abs/2111.13679

• Homepage: https://bmild.github.io/rawnerf/

• Demo: https://www.youtube.com/watch?v=JtBS4KBcKVc

Urban Radiance Fields

• Homepage: https://urban-radiance-fields.github.io/

• Paper: https://arxiv.org/abs/2111.14643

• Demo: https://youtu.be/qGlq5DZT6uc

Pix2NeRF: Unsupervised Conditional π-GAN for Single Image to Neural Radiance Fields Translation

• Paper: https://arxiv.org/abs/2202.13162

• Code: https://github.com/HexagonPrime/Pix2NeRF

HumanNeRF: Free-viewpoint Rendering of Moving People from Monocular Video

• Homepage: https://grail.cs.washington.edu/projects/humannerf/

• Paper: https://arxiv.org/abs/2201.04127

• Demo: https://youtu.be/GM-RoZEymmw

Visual Transformer

Backbone

MPViT : Multi-Path Vision Transformer for Dense Prediction

• Paper: https://arxiv.org/abs/2112.11010

• Code: https://github.com/youngwanLEE/MPViT

应用(Application)

Language-based Video Editing via Multi-Modal Multi-Level Transformer

• Paper: https://arxiv.org/abs/2104.01122

• Code: None

MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Video

• Paper: https://arxiv.org/abs/2203.00859

• Code: None

Embracing Single Stride 3D Object Detector with Sparse Transformer

• Paper: https://arxiv.org/abs/2112.06375

• Code: https://github.com/TuSimple/SST

Multi-class Token Transformer for Weakly Supervised Semantic Segmentation

• Paper: https://arxiv.org/abs/2203.02891

• Code: https://github.com/xulianuwa/MCTformer

Spatio-temporal Relation Modeling for Few-shot Action Recognition

• Paper: https://arxiv.org/abs/2112.05132

• Code: https://github.com/Anirudh257/strm

Mask-guided Spectral-wise Transformer for Efficient Hyperspectral Image Reconstruction

• Paper: https://arxiv.org/abs/2111.07910

• Code: https://github.com/caiyuanhao1998/MST

Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modeling

• Homepage: https://point-bert.ivg-research.xyz/

• Paper: https://arxiv.org/abs/2111.14819

• Code: https://github.com/lulutang0608/Point-BERT

GroupViT: Semantic Segmentation Emerges from Text Supervision

• Homepage: https://jerryxu.net/GroupViT/

• Paper: https://arxiv.org/abs/2202.11094

• Demo: https://youtu.be/DtJsWIUTW-Y

Restormer: Efficient Transformer for High-Resolution Image Restoration

• Paper: https://arxiv.org/abs/2111.09881

• Code: https://github.com/swz30/Restormer

Splicing ViT Features for Semantic Appearance Transfer

• Homepage: https://splice-vit.github.io/

• Paper: https://arxiv.org/abs/2201.00424

• Code: https://github.com/omerbt/Splice

Self-supervised Video Transformer

• Homepage: https://kahnchana.github.io/svt/

• Paper: https://arxiv.org/abs/2112.01514

• Code: https://github.com/kahnchana/svt

Learning Affinity from Attention: End-to-End Weakly-Supervised Semantic Segmentation with Transformers

• Paper: https://arxiv.org/abs/2203.02664

• Code: https://github.com/rulixiang/afa

Accelerating DETR Convergence via Semantic-Aligned Matching

• Paper: https://arxiv.org/abs/2203.06883

• Code: https://github.com/ZhangGongjie/SAM-DETR

DN-DETR: Accelerate DETR Training by Introducing Query DeNoising

• Paper: https://arxiv.org/abs/2203.01305

• Code: https://github.com/FengLi-ust/DN-DETR

• 中文解读: https://mp.weixin.qq.com/s/xdMfZ_L628Ru1d1iaMny0w

Style Transformer for Image Inversion and Editing

• Paper: https://arxiv.org/abs/2203.07932

• Code: https://github.com/sapphire497/style-transformer

MonoDTR: Monocular 3D Object Detection with Depth-Aware Transformer

• Paper: https://arxiv.org/abs/2203.10981

• Code: https://github.com/kuanchihhuang/MonoDTR

Mask Transfiner for High-Quality Instance Segmentation

• Paper: https://arxiv.org/abs/2111.13673

• Code: https://github.com/SysCV/transfiner

视觉和语言(Vision-Language)

Conditional Prompt Learning for Vision-Language Models

• Paper: https://arxiv.org/abs/2203.05557

• Code: https://github.com/KaiyangZhou/CoOp

自监督学习(Self-supervised Learning)

UniVIP: A Unified Framework for Self-Supervised Visual Pre-training

• Paper: https://arxiv.org/abs/2203.06965

• Code: None

Crafting Better Contrastive Views for Siamese Representation Learning

• Paper: https://arxiv.org/abs/2202.03278

• Code: https://github.com/xyupeng/ContrastiveCrop

• 中文解读：https://mp.weixin.qq.com/s/VTP9D5f7KG9vg30U9kVI2A

HCSC: Hierarchical Contrastive Selective Coding

• Homepage: https://github.com/gyfastas/HCSC

• Paper: https://arxiv.org/abs/2202.00455

• 中文解读: https://mp.weixin.qq.com/s/jkYR8mYp-e645qk8kfPNKQ

数据增强(Data Augmentation)

TeachAugment: Data Augmentation Optimization Using Teacher Knowledge

• Paper: https://arxiv.org/abs/2202.12513

• Code: https://github.com/DensoITLab/TeachAugment

AlignMix: Improving representation by interpolating aligned features

• Paper: https://arxiv.org/abs/2103.15375

• Code: None

目标检测(Object Detection)

DN-DETR: Accelerate DETR Training by Introducing Query DeNoising

• Paper: https://arxiv.org/abs/2203.01305

• Code: https://github.com/FengLi-ust/DN-DETR

• 中文解读: https://mp.weixin.qq.com/s/xdMfZ_L628Ru1d1iaMny0w

Accelerating DETR Convergence via Semantic-Aligned Matching

• Paper: https://arxiv.org/abs/2203.06883

• Code: https://github.com/ZhangGongjie/SAM-DETR

Localization Distillation for Dense Object Detection

• Paper: https://arxiv.org/abs/2102.12252

• Code: https://github.com/HikariTJU/LD

• Code2: https://github.com/HikariTJU/LD

• 中文解读：https://mp.weixin.qq.com/s/dxss8RjJH283h6IbPCT9vg

Focal and Global Knowledge Distillation for Detectors

• Paper: https://arxiv.org/abs/2111.11837

• Code: https://github.com/yzd-v/FGD

• 中文解读：​​https://mp.weixin.qq.com/s/yDkreTudC8JL2V2ETsADwQ​​

A Dual Weighting Label Assignment Scheme for Object Detection

• Paper: https://arxiv.org/abs/2203.09730

• Code: https://github.com/strongwolf/DW

目标跟踪(Visual Tracking)

Correlation-Aware Deep Tracking

• Paper: https://arxiv.org/abs/2203.01666

• Code: None

TCTrack: Temporal Contexts for Aerial Tracking

• Paper: https://arxiv.org/abs/2203.01885

• Code: https://github.com/vision4robotics/TCTrack

语义分割(Semantic Segmentation)

弱监督语义分割

Class Re-Activation Maps for Weakly-Supervised Semantic Segmentation

• Paper: https://arxiv.org/abs/2203.00962

• Code: https://github.com/zhaozhengChen/ReCAM

Multi-class Token Transformer for Weakly Supervised Semantic Segmentation

• Paper: https://arxiv.org/abs/2203.02891

• Code: https://github.com/xulianuwa/MCTformer

Learning Affinity from Attention: End-to-End Weakly-Supervised Semantic Segmentation with Transformers

• Paper: https://arxiv.org/abs/2203.02664

• Code: https://github.com/rulixiang/afa

半监督语义分割

ST++: Make Self-training Work Better for Semi-supervised Semantic Segmentation

• Paper: https://arxiv.org/abs/2106.05095

• Code: https://github.com/LiheYoung/ST-PlusPlus

• 中文解读：​​https://mp.weixin.qq.com/s/knSnlebdtEnmrkChGM_0CA​​

Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels

• Homepage: https://haochen-wang409.github.io/U2PL/

• Paper: https://arxiv.org/abs/2203.03884

• Code: https://github.com/Haochen-Wang409/U2PL

• 中文解读: ​​https://mp.weixin.qq.com/s/-08olqE7np8A1XQzt6HAgQ​​

无监督语义分割

GroupViT: Semantic Segmentation Emerges from Text Supervision

• Homepage: https://jerryxu.net/GroupViT/

• Paper: https://arxiv.org/abs/2202.11094

• Demo: https://youtu.be/DtJsWIUTW-Y

实例分割(Instance Segmentation)

E2EC: An End-to-End Contour-based Method for High-Quality High-Speed Instance Segmentation

• Paper: https://arxiv.org/abs/2203.04074

• Code: https://github.com/zhang-tao-whu/e2ec

Mask Transfiner for High-Quality Instance Segmentation

• Paper: https://arxiv.org/abs/2111.13673

• Code: https://github.com/SysCV/transfiner

自监督实例分割

FreeSOLO: Learning to Segment Objects without Annotations

• Paper: https://arxiv.org/abs/2202.12181

• Code: None

视频实例分割

Efficient Video Instance Segmentation via Tracklet Query and Proposal

• Homepage: https://jialianwu.com/projects/EfficientVIS.html

• Paper: https://arxiv.org/abs/2203.01853

• Demo: https://youtu.be/sSPMzgtMKCE

小样本分割(Few-Shot Segmentation)

Learning What Not to Segment: A New Perspective on Few-Shot Segmentation

• Paper: https://arxiv.org/abs/2203.07615

• Code: https://github.com/chunbolang/BAM

视频理解(Video Understanding)

Self-supervised Video Transformer

• Homepage: https://kahnchana.github.io/svt/

• Paper: https://arxiv.org/abs/2112.01514

• Code: https://github.com/kahnchana/svt

行为识别(Action Recognition)

Spatio-temporal Relation Modeling for Few-shot Action Recognition

• Paper: https://arxiv.org/abs/2112.05132

• Code: https://github.com/Anirudh257/strm

动作检测(Action Detection)

End-to-End Semi-Supervised Learning for Video Action Detection

• Paper: https://arxiv.org/abs/2203.04251

• Code: None

图像编辑(Image Editing)

Style Transformer for Image Inversion and Editing

• Paper: https://arxiv.org/abs/2203.07932

• Code: https://github.com/sapphire497/style-transformer

Blended Diffusion for Text-driven Editing of Natural Images

• Paper: https://arxiv.org/abs/2111.14818

• Code: https://github.com/omriav/blended-diffusion

SemanticStyleGAN: Learning Compositional Generative Priors for Controllable Image Synthesis and Editing

• Homepage: https://semanticstylegan.github.io/

• Paper: https://arxiv.org/abs/2112.02236

• Demo: https://semanticstylegan.github.io/videos/demo.mp4

Low-level Vision

ISNAS-DIP: Image-Specific Neural Architecture Search for Deep Image Prior

• Paper: https://arxiv.org/abs/2111.15362

• Code: None

Restormer: Efficient Transformer for High-Resolution Image Restoration

• Paper: https://arxiv.org/abs/2111.09881

• Code: https://github.com/swz30/Restormer

超分辨率(Super-Resolution)

图像超分辨率(Image Super-Resolution)

Learning the Degradation Distribution for Blind Image Super-Resolution

• Paper: https://arxiv.org/abs/2203.04962

• Code: https://github.com/greatlog/UnpairedSR

视频超分辨率(Video Super-Resolution)

BasicVSR++: Improving Video Super-Resolution with Enhanced Propagation and Alignment

• Paper: https://arxiv.org/abs/2104.13371

• Code: https://github.com/open-mmlab/mmediting

• Code: https://github.com/ckkelvinchan/BasicVSR_PlusPlus

• 中文解读：​​https://mp.weixin.qq.com/s/HZTwYfphixyLHxlbCAxx4g​​

3D点云(3D Point Cloud)

Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modeling

• Homepage: https://point-bert.ivg-research.xyz/

• Paper: https://arxiv.org/abs/2111.14819

• Code: https://github.com/lulutang0608/Point-BERT

A Unified Query-based Paradigm for Point Cloud Understanding

• Paper: https://arxiv.org/abs/2203.01252

• Code: None

CrossPoint: Self-Supervised Cross-Modal Contrastive Learning for 3D Point Cloud Understanding

• Paper: https://arxiv.org/abs/2203.00680

• Code: https://github.com/MohamedAfham/CrossPoint

PointCLIP: Point Cloud Understanding by CLIP

• Paper: https://arxiv.org/abs/2112.02413

• Code: https://github.com/ZrrSkywalker/PointCLIP

3D目标检测(3D Object Detection)

Embracing Single Stride 3D Object Detector with Sparse Transformer

• Paper: https://arxiv.org/abs/2112.06375

• Code: https://github.com/TuSimple/SST

Canonical Voting: Towards Robust Oriented Bounding Box Detection in 3D Scenes

• Paper: https://arxiv.org/abs/2011.12001

• Code: https://github.com/qq456cvb/CanonicalVoting

MonoDTR: Monocular 3D Object Detection with Depth-Aware Transformer

• Paper: https://arxiv.org/abs/2203.10981

• Code: https://github.com/kuanchihhuang/MonoDTR

3D语义分割(3D Semantic Segmentation)

Scribble-Supervised LiDAR Semantic Segmentation

• Paper: https://arxiv.org/abs/2203.08537

• Dataset: https://github.com/ouenal/scribblekitti

3D目标跟踪(3D Object Tracking)

Beyond 3D Siamese Tracking: A Motion-Centric Paradigm for 3D Single Object Tracking in Point Clouds

• Paper: https://arxiv.org/abs/2203.01730

• Code: https://github.com/Ghostish/Open3DSOT

3D人体姿态估计(3D Human Pose Estimation)

MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimation

• Paper: https://arxiv.org/abs/2111.12707

• Code: https://github.com/Vegetebird/MHFormer

• 中文解读: https://zhuanlan.zhihu.com/p/439459426

MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Video

• Paper: https://arxiv.org/abs/2203.00859

• Code: None

3D语义场景补全(3D Semantic Scene Completion)

MonoScene: Monocular 3D Semantic Scene Completion

• Paper: https://arxiv.org/abs/2112.00726

• Code: https://github.com/cv-rits/MonoScene

3D重建(3D Reconstruction)

BANMo: Building Animatable 3D Neural Models from Many Casual Videos

• Homepage: https://banmo-www.github.io/

• Paper: https://arxiv.org/abs/2112.12761

• Code: https://github.com/facebookresearch/banmo

• 中文解读：​​https://mp.weixin.qq.com/s/NMHP8-xWwrX40vpGx55Qew​​

伪装物体检测(Camouflaged Object Detection)

Zoom In and Out: A Mixed-scale Triplet Network for Camouflaged Object Detection

• Paper: https://arxiv.org/abs/2203.02688

• Code: https://github.com/lartpang/ZoomNet

深度估计(Depth Estimation)

单目深度估计

NeW CRFs: Neural Window Fully-connected CRFs for Monocular Depth Estimation

• Paper: https://arxiv.org/abs/2203.01502

• Code: None

OmniFusion: 360 Monocular Depth Estimation via Geometry-Aware Fusion

• Paper: https://arxiv.org/abs/2203.00838

• Code: None

Toward Practical Self-Supervised Monocular Indoor Depth Estimation

• Paper: https://arxiv.org/abs/2112.02306

• Code: None

立体匹配(Stereo Matching)

ACVNet: Attention Concatenation Volume for Accurate and Efficient Stereo Matching

• Paper: https://arxiv.org/abs/2203.02146

• Code: https://github.com/gangweiX/ACVNet

车道线检测(Lane Detection)

Rethinking Efficient Lane Detection via Curve Modeling

• Paper: https://arxiv.org/abs/2203.02431

• Code: https://github.com/voldemortX/pytorch-auto-drive

• Demo：https://user-images.githubusercontent.com/32259501/148680744-a18793cd-f437-461f-8c3a-b909c9931709.mp4

图像修复(Image Inpainting)

Incremental Transformer Structure Enhanced Image Inpainting with Masking Positional Encoding

• Paper: https://arxiv.org/abs/2203.00867

• Code: https://github.com/DQiaole/ZITS_inpainting

人群计数(Crowd Counting)

Leveraging Self-Supervision for Cross-Domain Crowd Counting

• Paper: https://arxiv.org/abs/2103.16291

• Code: None

医学图像(Medical Image)

BoostMIS: Boosting Medical Image Semi-supervised Learning with Adaptive Pseudo Labeling and Informative Active Annotation

• Paper: https://arxiv.org/abs/2203.02533

• Code: None

场景图生成(Scene Graph Generation)

SGTR: End-to-end Scene Graph Generation with Transformer

• Paper: https://arxiv.org/abs/2112.12970

• Code: None

风格迁移(Style Transfer)

StyleMesh: Style Transfer for Indoor 3D Scene Reconstructions

• Homepage: https://lukashoel.github.io/stylemesh/

• Paper: https://arxiv.org/abs/2112.01530

• Code: https://github.com/lukasHoel/stylemesh

• Demo：https://www.youtube.com/watch?v=ZqgiTLcNcks

弱监督物体检测(Weakly Supervised Object Localization)

Weakly Supervised Object Localization as Domain Adaption

• Paper: https://arxiv.org/abs/2203.01714

• Code: https://github.com/zh460045050/DA-WSOL_CVPR2022

高光谱图像重建(Hyperspectral Image Reconstruction)

Mask-guided Spectral-wise Transformer for Efficient Hyperspectral Image Reconstruction

• Paper: https://arxiv.org/abs/2111.07910

• Code: https://github.com/caiyuanhao1998/MST

水印(Watermarking)

Deep 3D-to-2D Watermarking: Embedding Messages in 3D Meshes and Extracting Them from 2D Renderings

• Paper: https://arxiv.org/abs/2104.13450

• Code: None

数据集(Datasets)

It's About Time: Analog Clock Reading in the Wild

• Homepage: https://charigyang.github.io/abouttime/

• Paper: https://arxiv.org/abs/2111.09162

• Code: https://github.com/charigyang/itsabouttime

• Demo: https://youtu.be/cbiMACA6dRc

Toward Practical Self-Supervised Monocular Indoor Depth Estimation

• Paper: https://arxiv.org/abs/2112.02306

• Code: None

Kubric: A scalable dataset generator

• Paper: https://arxiv.org/abs/2203.03570

• Code: https://github.com/google-research/kubric

Scribble-Supervised LiDAR Semantic Segmentation

• Paper: https://arxiv.org/abs/2203.08537

• Dataset: https://github.com/ouenal/scribblekitti

新任务(New Task)

Language-based Video Editing via Multi-Modal Multi-Level Transformer

• Paper: https://arxiv.org/abs/2104.01122

• Code: None

It's About Time: Analog Clock Reading in the Wild

• Homepage: https://charigyang.github.io/abouttime/

• Paper: https://arxiv.org/abs/2111.09162

• Code: https://github.com/charigyang/itsabouttime

• Demo: https://youtu.be/cbiMACA6dRc

Splicing ViT Features for Semantic Appearance Transfer

• Homepage: https://splice-vit.github.io/

• Paper: https://arxiv.org/abs/2201.00424

• Code: https://github.com/omerbt/Splice

其他(Others)

Kubric: A scalable dataset generator

• Paper: https://arxiv.org/abs/2203.03570

• Code: https://github.com/google-research/kubric

​机器学习算法AI大数据技术​

​ 搜索公众号添加： datanlp​

​长按图片，识别二维码​

​​

​不断更新资源​

​深度学习、机器学习、数据分析、python​

​ 搜索公众号添加： datayx  ​