ECCV2020-Code

ECCV 2020 论文开源项目合集，同时欢迎各位大佬提交issue，分享ECCV 2020开源项目

关于往年CV顶会论文（如CVPR 2020、ICCV 2019、ECCV 2018）以及其他优质CV论文和大盘点，详见： https://github.com/amusi/daily-paper-computer-vision

• CNN
• 图像分类
• 2D目标检测
• 3D目标检测
• 视频目标检测
• 语义分割
• 实例分割
• 全景分割
• 视频目标分割
• 单/多目标跟踪
• GAN
• NAS
• 3D点云（分类/分割/配准/补全等）
• 人脸（检测/识别/解析等）
• Re-ID
• 显著性检测（SOD）
• 模型压缩（剪枝/知识蒸馏等）
• 视频理解/行为识别/行为检测
• 场景文本检测
• 场景文本识别
• 特征点检测/描述符/匹配
• 姿态估计
• 深度估计
• 深度补全
• 域泛化
• 超分辨率
• 去模糊
• 去雨
• 图像/视频恢复
• 图像/视频修复(补全)
• 风格迁移
• 三维重建
• 图像描述
• 图像检索
• 光流估计
• 视频插帧
• 车道线检测
• 轨迹预测
• 线段检测
• 视线估计
• 眼动追踪
• 对抗攻击
• 数据集
• 其他
• 不确定中没中

CNN

Beyond Fixed Grid: Learning Geometric Image Representation with a Deformable Grid

• 主页：http://www.cs.toronto.edu/~jungao/def-grid/
• 论文：http://xxx.itp.ac.cn/abs/2008.09269
• 代码：https://github.com/fidler-lab/deformable-grid-release

WeightNet: Revisiting the Design Space of Weight Networks

• 论文：https://arxiv.org/abs/2007.11823
• 代码：https://github.com/megvii-model/WeightNet

Feature Pyramid Transformer

• 论文：https://arxiv.org/abs/2007.09451

• 代码：https://github.com/ZHANGDONG-NJUST/FPT

Dynamic Group Convolution for Accelerating Convolutional Neural Networks

• 论文：https://arxiv.org/abs/2007.04242
• 代码：https://github.com/zhuogege1943/dgc

Learning to Learn Parameterized Classification Networks for Scalable Input Images

• 论文：https://arxiv.org/abs/2007.06181

• 代码：https://github.com/d-li14/SAN

Rethinking Bottleneck Structure for Efficient Mobile Network Design

• 论文：https://arxiv.org/abs/2007.02269
• 代码：https://github.com/zhoudaquan/rethinking_bottleneck_design

MutualNet: Adaptive ConvNet via Mutual Learning from Network Width and Resolution

• 论文：Oral
• 论文：https://arxiv.org/abs/1909.12978
• 代码：https://github.com/taoyang1122/MutualNet

PSConv: Squeezing Feature Pyramid into One Compact Poly-Scale Convolutional Layer

• 论文：https://arxiv.org/abs/2007.06191
• 代码：https://github.com/d-li14/PSConv

图像分类

Learning to Learn Parameterized Classification Networks for Scalable Input Images

• 论文：暂无

• 代码：https://github.com/d-li14/SAN

Learning To Classify Images Without Labels

• 论文：https://arxiv.org/abs/2005.12320
• 代码：https://github.com/wvangansbeke/Unsupervised-Classification

2D目标检测

Learning Data Augmentation Strategies for Object Detection

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/5865_ECCV_2020_paper.php
• 代码：https://github.com/tensorflow/tpu/tree/master/models/official/detection

AABO: Adaptive Anchor Box Optimization for Object Detection via Bayesian Sub-sampling

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3977_ECCV_2020_paper.php
• 代码：https://github.com/wwdkl/AABO

Side-Aware Boundary Localization for More Precise Object Detection

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2272_ECCV_2020_paper.php
• 代码：https://github.com/open-mmlab/mmdetection

TIDE: A General Toolbox for Identifying Object Detection Errors

• 主页：https://dbolya.github.io/tide/

• 论文：https://arxiv.org/abs/2008.08115

• 代码：https://github.com/dbolya/tide

Every Pixel Matters: Center-aware Feature Alignment for Domain Adaptive Object Detector

• 主页：https://chengchunhsu.github.io/EveryPixelMatters/
• 论文：https://arxiv.org/abs/2008.08574
• 代码：https://github.com/chengchunhsu/EveryPixelMatters

Dense RepPoints: Representing Visual Objects with Dense Point Sets

• 论文：https://arxiv.org/abs/1912.11473
• 代码：https://github.com/justimyhxu/Dense-RepPoints

Corner Proposal Network for Anchor-free, Two-stage Object Detection

• 论文：https://arxiv.org/abs/2007.13816

• 代码：https://github.com/Duankaiwen/CPNDet

BorderDet: Border Feature for Dense Object Detection

• 论文：https://arxiv.org/abs/2007.11056

• 代码：https://github.com/Megvii-BaseDetection/BorderDet

• 中文解读：https://zhuanlan.zhihu.com/p/163044323

Multi-Scale Positive Sample Refinement for Few-Shot Object Detection

• 论文：https://arxiv.org/abs/2007.09384

• 代码：https://github.com/jiaxi-wu/MPSR

PIoU Loss: Towards Accurate Oriented Object Detection in Complex Environments

• 论文：https://arxiv.org/abs/2007.09584

• 代码：https://github.com/clobotics/piou

• 数据集：https://github.com/clobotics/piou

Probabilistic Anchor Assignment with IoU Prediction for Object Detection

• 论文：https://arxiv.org/abs/2007.08103
• 代码：https://github.com/kkhoot/PAA

HoughNet: Integrating near and long-range evidence for bottom-up object detection

• 论文：https://arxiv.org/abs/2007.02355
• 代码：https://github.com/nerminsamet/houghnet

OS2D: One-Stage One-Shot Object Detection by Matching Anchor Features

• 论文：https://arxiv.org/abs/2003.06800

• 代码：https://github.com/aosokin/os2d

End-to-End Object Detection with Transformers

• Oral

• 论文：https://ai.facebook.com/research/publications/end-to-end-object-detection-with-transformers

• 代码：https://github.com/facebookresearch/detr

Dynamic R-CNN: Towards High Quality Object Detection via Dynamic Training

• 论文：https://arxiv.org/abs/2004.06002
• 代码：https://github.com/hkzhang95/DynamicRCNN

OS2D: One-Stage One-Shot Object Detection by Matching Anchor Features

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2424_ECCV_2020_paper.php
• 代码：https://github.com/aosokin/os2d

Object Detection with a Unified Label Space from Multiple Datasets

• 主页：http://www.nec-labs.com/~mas/UniDet/
• 论文：https://arxiv.org/abs/2008.06614
• 代码：暂无
• 数据集：http://www.nec-labs.com/~mas/UniDet/resources/UOD_dataset_ECCV20.zip

弱监督目标检测

Enabling Deep Residual Networks for Weakly Supervised Object Detection

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/479_ECCV_2020_paper.php
• 代码：https://github.com/shenyunhang/DRN-WSOD

UFO²: A Unified Framework towards Omni-supervised Object Detection

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3205_ECCV_2020_paper.php
• 代码：https://github.com/NVlabs/wetectron

Boosting Weakly Supervised Object Detection with Progressive Knowledge Transfer

• 论文：https://arxiv.org/abs/2007.07986
• 代码：https://github.com/mikuhatsune/wsod_transfer

域自适应目标检测

Collaborative Training between Region Proposal Localization and Classification for Domain Adaptive Object Detection

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2920_ECCV_2020_paper.php
• 代码：https://github.com/GanlongZhao/CST_DA_detection

Every Pixel Matters: Center-aware Feature Alignment for Domain Adaptive Object Detector

• 主页：https://chengchunhsu.github.io/EveryPixelMatters/
• 论文：https://arxiv.org/abs/2008.08574
• 代码：https://github.com/chengchunhsu/EveryPixelMatters

Few-Shot 目标检测

Multi-Scale Positive Sample Refinement for Few-Shot Object Detection

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2581_ECCV_2020_paper.php
• 代码：https://github.com/jiaxi-wu/MPSR

水下目标检测

Dual Refinement Underwater Object Detection Network

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3425_ECCV_2020_paper.php
• 代码：https://github.com/Peterchen111/FERNet

遥感旋转目标检测

PIoU Loss: Towards Accurate Oriented Object Detection in Complex Environments

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3087_ECCV_2020_paper.php
• 代码：https://github.com/clobotics/piou
• 数据集：https://github.com/clobotics/piou

Arbitrary-Oriented Object Detection with Circular Smooth Label

• 论文：https://arxiv.org/abs/2003.05597
• 代码：https://github.com/Thinklab-SJTU/CSL_RetinaNet_Tensorflow

3D目标检测

Rethinking Pseudo-LiDAR Representation

• 论文：https://arxiv.org/abs/2008.04582

• 代码：https://github.com/xinzhuma/patchnet

Pillar-based Object Detection for Autonomous Driving

• 论文：https://arxiv.org/abs/2007.10323
• 代码：https://github.com/WangYueFt/pillar-od

EPNet: Enhancing Point Features with Image Semantics for 3D Object Detection

• 论文：https://arxiv.org/abs/2007.08856
• 代码：https://github.com/happinesslz/EPNet

视频目标检测

Mining Inter-Video Proposal Relations for Video Object Detection

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3764_ECCV_2020_paper.php
• 代码：https://github.com/youthHan/HVRNet

Learning Where to Focus for Efficient Video Object Detection

• 主页：https://jiangzhengkai.github.io/LSTS/
• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123610018.pdf
• 代码：https://github.com/jiangzhengkai/LSTS

语义分割

SNE-RoadSeg: Incorporating Surface Normal Information into Semantic Segmentation for Accurate Freespace Detection

• 主页：https://sites.google.com/view/sne-roadseg
• 论文：https://arxiv.org/abs/2008.11351
• 代码：https://github.com/hlwang1124/SNE-RoadSeg
• 数据集：https://drive.google.com/file/d/1YnkqPmzxtjNfMi2B1gMy7LQa5Gnu-BsH/view

Tensor Low-Rank Reconstruction for Semantic Segmentation

• 论文：https://arxiv.org/abs/2008.00490

• 代码：https://github.com/CWanli/RecoNet

Bi-directional Cross-Modality Feature Propagation with Separation-and-Aggregation Gate for RGB-D Semantic Segmentation

• 论文：https://arxiv.org/abs/2007.09183

• 代码：https://github.com/charlesCXK/RGBD_Semantic_Segmentation_PyTorch

GMNet: Graph Matching Network for Large Scale Part Semantic Segmentation in the Wild

• 主页：https://lttm.dei.unipd.it/paper_data/GMNet/
• 论文：https://arxiv.org/abs/2007.09073
• 代码：https://github.com/LTTM/GMNet

SegFix: Model-Agnostic Boundary Refinement for Segmentation

• 论文：https://arxiv.org/abs/2007.04269

• 代码：https://github.com/openseg-group/openseg.pytorch

Mining Cross-Image Semantics for Weakly Supervised Semantic Segmentation

• Oral
• 论文：https://arxiv.org/abs/2007.01947
• 代码：https://github.com/GuoleiSun/MCIS_wsss

Improving Semantic Segmentation via Decoupled Body and Edge Supervision

• 论文：https://arxiv.org/abs/2007.10035

• 代码：https://github.com/lxtGH/DecoupleSegNets

实例分割

SipMask: Spatial Information Preservation for Fast Image and Video Instance Segmentation

• 论文：https://arxiv.org/abs/2007.14772

• 代码：https://github.com/JialeCao001/SipMask

Commonality-Parsing Network across Shape and Appearance for Partially Supervised Instance Segmentation

• 论文：https://arxiv.org/abs/2007.12387

• 代码：https://github.com/fanq15/CPMask

Boundary-preserving Mask R-CNN

• 论文：https://arxiv.org/abs/2007.08921

• 代码：https://github.com/hustvl/BMaskR-CNN

Conditional Convolutions for Instance Segmentation

• Oral
• 论文：https://arxiv.org/abs/2003.05664
• 代码：https://github.com/aim-uofa/AdelaiDet/blob/master/configs/CondInst/README.md

SOLO: Segmenting Objects by Locations

• 论文：https://arxiv.org/abs/1912.04488

• 代码：https://github.com/WXinlong/SOLO

• 知乎话题：https://www.zhihu.com/question/360594484

全景分割

Axial-DeepLab: Stand-Alone Axial-Attention for Panoptic Segmentation

• 论文：https://arxiv.org/abs/2003.07853

• 代码：https://github.com/csrhddlam/axial-deeplab

• 视频：https://youtu.be/-iAXF-vibdE

视频目标分割

Collaborative Video Object Segmentation by Foreground-Background Integration

• 论文：https://arxiv.org/abs/2003.08333
• 代码：https://github.com/z-x-yang/CFBI

Video Object Segmentation with Episodic Graph Memory Networks

• 论文：https://arxiv.org/abs/2007.07020

• 代码：https://github.com/carrierlxk/GraphMemVOS

单/多目标跟踪

Ocean: Object-aware Anchor-Free Tracking

• 论文：https://arxiv.org/abs/2006.10721

• 代码：https://github.com/researchmm/TracKit

多目标跟踪

Towards Real-Time Multi-Object Tracking

• 论文：暂无
• 代码：https://github.com/Zhongdao/Towards-Realtime-MOT

Simultaneous Detection and Tracking with Motion Modelling for Multiple Object Tracking

• 论文：https://arxiv.org/abs/2008.08826
• 代码：https://github.com/shijieS/DMMN
• 数据集：https://github.com/shijieS/OmniMOTDataset

Chained-Tracker: Chaining Paired Attentive Regression Results for End-to-End Joint Multiple-Object Detection and Tracking

• 论文：https://arxiv.org/abs/2007.14557

• 代码：https://github.com/pjl1995/CTracker

Ocean: Object-aware Anchor-Free Tracking

• 论文：https://arxiv.org/abs/2006.10721

• 代码：https://github.com/researchmm/TracKit

TAO: A Large-Scale Benchmark for Tracking Any Object

• 主页：http://taodataset.org/
• 论文：https://arxiv.org/abs/2005.10356
• 代码：https://github.com/TAO-Dataset/tao

Segment as Points for Efficient Online Multi-Object Tracking and Segmentation

• Oral
• 论文：https://arxiv.org/abs/2007.01550
• 代码：https://github.com/detectRecog/PointTrack
• 数据集：https://github.com/detectRecog/PointTrack

GAN

Rewriting a Deep Generative Model

• 论文：https://arxiv.org/abs/2007.15646

• 代码：https://github.com/davidbau/rewriting

Contrastive Learning for Unpaired Image-to-Image Translation

• 论文：https://arxiv.org/abs/2007.15651
• 代码：https://github.com/taesungp/contrastive-unpaired-translation

XingGAN for Person Image Generation

• 论文：暂无
• 代码：https://github.com/Ha0Tang/XingGAN

NAS

Are Labels Necessary for Neural Architecture Search?

• 论文：https://arxiv.org/abs/2003.12056

• 代码：https://github.com/facebookresearch/unnas

Rethinking Bottleneck Structure for Efficient Mobile Network Design

• 论文：https://arxiv.org/abs/2007.02269
• 代码：https://github.com/zhoudaquan/rethinking_bottleneck_design

Fair DARTS: Eliminating Unfair Advantages in Differentiable Architecture Search

• 论文：https://arxiv.org/abs/1911.12126
• 代码：https://github.com/xiaomi-automl/fairdarts

3D点云（分类/分割/配准/补全等）

AdvPC: Transferable Adversarial Perturbations on 3D Point Clouds

• 论文：https://arxiv.org/abs/1912.00461

• 代码：https://github.com/ajhamdi/AdvPC

A Closer Look at Local Aggregation Operators in Point Cloud Analysis

• 论文：https://arxiv.org/abs/2007.01294
• 代码：https://github.com/zeliu98/CloserLook3D

3D点云补全

Multimodal Shape Completion via Conditional Generative Adversarial Networks

• 论文：https://arxiv.org/abs/2003.07717
• 代码：https://github.com/ChrisWu1997/Multimodal-Shape-Completion

GRNet: Gridding Residual Network for Dense Point Cloud Completion

• 论文：https://arxiv.org/abs/2006.03761
• 代码：https://github.com/hzxie/GRNet

3D点云生成

Progressive Point Cloud Deconvolution Generation Network

• 论文：https://arxiv.org/abs/2007.05361

• 代码：https://github.com/fpthink/PDGN

人脸（检测/识别/解析等）

人脸检测

ProgressFace: Scale-Aware Progressive Learning for Face Detection

• 论文：http://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123510341.pdf
• 代码：https://github.com/jiashu-zhu/ProgressFace

人脸识别

Explainable Face Recognition

• 论文：https://arxiv.org/abs/2008.00916

• 主页：https://stresearch.github.io/xfr/

• 代码：https://github.com/stresearch/xfr

3D人脸重建

Self-Supervised Monocular 3D Face Reconstruction by Occlusion-Aware Multi-view Geometry Consistency

• 论文：https://arxiv.org/abs/2007.12494
• 代码：https://github.com/jiaxiangshang/MGCNet

人脸活体检测

CelebA-Spoof: Large-Scale Face Anti-Spoofing Dataset with Rich Annotations

• 论文：https://arxiv.org/abs/2007.12342

• 数据集：https://github.com/Davidzhangyuanhan/CelebA-Spoof

人脸解析

Edge-aware Graph Representation Learning and Reasoning for Face Parsing

• 论文：https://arxiv.org/abs/2007.11240
• 代码：https://github.com/tegusi/EAGRNet

DeepFakes

What makes fake images detectable? Understanding properties that generalize

• 主页：https://chail.github.io/patch-forensics/
• 论文：http://xxx.itp.ac.cn/abs/2008.10588
• 代码：https://github.com/chail/patch-forensics

其他

Lifespan Age Transformation Synthesis

• 论文：https://arxiv.org/abs/2003.09764
• 代码：https://github.com/royorel/Lifespan_Age_Transformation_Synthesis

Re-ID

行人重识别

Rethinking the Distribution Gap of Person Re-identification with Camera-based Batch Normalization

• 论文：https://arxiv.org/abs/2001.08680
• 代码：https://github.com/automan000/Camera-based-Person-ReID

Appearance-Preserving 3D Convolution for Video-based Person Re-identification

• Oral

• 论文：https://arxiv.org/pdf/2007.08434

• 代码：https://github.com/guxinqian/AP3D

Do Not Disturb Me: Person Re-identification Under the Interference of Other Pedestrians

• 论文：https://arxiv.org/abs/2008.06963
• 代码：https://github.com/X-BrainLab/PI-ReID

Faster Person Re-Identification

• 论文：https://arxiv.org/abs/2008.06826

• 代码：https://github.com/wangguanan/light-reid

Temporal Complementary Learning for Video Person Re-Identification

• 论文：https://arxiv.org/abs/2007.09357

• 代码：https://github.com/blue-blue272/VideoReID-TCLNet

Joint Disentangling and Adaptation for Cross-Domain Person Re-Identification

• 论文：https://arxiv.org/abs/2007.10315
• 代码：https://github.com/NVlabs/DG-Net-PP

Robust Re-Identification by Multiple Views Knowledge Distillation

• 论文：https://arxiv.org/abs/2007.04174
• 代码：https://github.com/aimagelab/VKD

Multiple Expert Brainstorming for Domain Adaptive Person Re-identification

• 论文：https://arxiv.org/abs/2007.01546
• 代码：https://github.com/YunpengZhai/MEB-Net

车辆重识别

Simulating Content Consistent Vehicle Datasets with Attribute Descent

• 论文：https://arxiv.org/abs/1912.08855
• 代码：https://github.com/yorkeyao/VehicleX
• 数据集：https://github.com/yorkeyao/VehicleX

Orientation-aware Vehicle Re-identification with Semantics-guided Part Attention Network

• 主页：http://media.ee.ntu.edu.tw/research/SPAN/

• 论文：https://arxiv.org/abs/2008.11423

• 代码：https://github.com/tsaishien-chen/SPAN

显著性检测（SOD）

Progressively Guided Alternate Refinement Network for RGB-D Salient Object Detection

• 论文：http://xxx.itp.ac.cn/abs/2008.07064

• 代码：https://github.com/ShuhanChen/PGAR_ECCV20

Suppress and Balance: A Simple Gated Network for Salient Object Detection

• Oral

• 论文：https://arxiv.org/abs/2007.08074

• 代码：https://github.com/Xiaoqi-Zhao-DLUT/GateNet-RGB-Saliency

Hierarchical Dynamic Filtering Network for RGB-D Salient Object Detection

• 论文：https://arxiv.org/abs/2007.06227

• 代码：https://github.com/lartpang/HDFNet

A Single Stream Network for Robust and Real-time RGB-D Salient Object Detection

• 论文：https://arxiv.org/abs/2007.06811
• 代码：https://github.com/Xiaoqi-Zhao-DLUT/DANet-RGBD-Saliency

Cross-Modal Weighting Network for RGB-D Salient Object Detection

• 论文：暂无

• 代码：https://github.com/MathLee/CMWNet

BBS-Net: RGB-D Salient Object Detection with a Bifurcated Backbone Strategy Network

• 论文：暂无
• 代码：https://github.com/DengPingFan/BBS-Net

Highly Efficient Salient Object Detection with 100K Parameters

• 论文：https://arxiv.org/abs/2003.05643
• 代码：https://github.com/MCG-NKU/Sal100K

模型压缩（剪枝/知识蒸馏等）

EagleEye: Fast Sub-net Evaluation for Efficient Neural Network Pruning

• 论文：https://arxiv.org/abs/2007.02491
• 代码：https://github.com/anonymous47823493/EagleEye

视频理解/行为识别/行为检测

AssembleNet++: Assembling Modality Representations via Attention Connections

• 论文：https://arxiv.org/abs/2008.08072
• 代码：https://sites.google.com/corp/view/assemblenet/

LEMMA: A Multi-view Dataset for Learning Multi-agent Multi-task Activities

• 主页：https://sites.google.com/view/lemma-activity

• 论文：https://arxiv.org/abs/2007.15781

• 数据集：https://sites.google.com/view/lemma-activity/home/dataset

• 代码：https://github.com/Buzz-Beater/LEMMA

AR-Net: Adaptive Frame Resolution for Efficient Action Recognition

• 主页：https://mengyuest.github.io/AR-Net/
• 论文：https://arxiv.org/abs/2007.15796
• 代码：https://github.com/mengyuest/AR-Net

Context-Aware RCNN: A Baseline for Action Detection in Videos

• 论文：https://arxiv.org/abs/2007.09861

• 代码：https://github.com/MCG-NJU/CRCNN-Action

Actions as Moving Points

• 论文：https://arxiv.org/abs/2001.04608
• 代码：https://github.com/MCG-NJU/MOC-Detector

SF-Net: Single-Frame Supervision for Temporal Action Localization

• 论文：https://arxiv.org/abs/2003.06845
• 代码：https://github.com/Flowerfan/SF-Net

Asynchronous Interaction Aggregation for Action Detection

• 论文：https://arxiv.org/abs/2004.07485

• 代码：https://github.com/MVIG-SJTU/AlphAction

场景文本检测

Mask TextSpotter v3: Segmentation Proposal Network for Robust Scene Text Spotting

• 论文：https://arxiv.org/abs/2007.09482

• 代码：https://github.com/MhLiao/MaskTextSpotterV3

场景文本识别

Adaptive Text Recognition through Visual Matching

• 主页：http://www.robots.ox.ac.uk/~vgg/research/FontAdaptor20/

• 论文：https://arxiv.org/abs/2009.06610

• 代码：https://github.com/Chuhanxx/FontAdaptor

Mask TextSpotter v3: Segmentation Proposal Network for Robust Scene Text Spotting

• 论文：https://arxiv.org/abs/2007.09482

• 代码：https://github.com/MhLiao/MaskTextSpotterV3

特征点检测/描述符/匹配

Learning and aggregating deep local descriptors for instance-level recognition

• 论文：https://arxiv.org/abs/2007.13172

• 代码：https://github.com/gtolias/how

Online Invariance Selection for Local Feature Descriptors

• Oral
• 论文：https://arxiv.org/abs/2007.08988
• 代码：https://github.com/rpautrat/LISRD

Single-Image Depth Prediction Makes Feature Matching Easier

• 论文：https://arxiv.org/abs/2008.09497

• 代码：http://www.github.com/nianticlabs/rectified-features

姿态估计

Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose

• 论文：https://arxiv.org/abs/2008.09047
• 代码：https://github.com/hongsukchoi/Pose2Mesh_RELEASE

Key Frame Proposal Network for Efficient Pose Estimation in Videos

• 论文：https://arxiv.org/abs/2007.15217
• 代码：https://github.com/Yuexiaoxi10/Key-Frame-Proposal-Network-for-Efficient-Pose-Estimation-in-Videos

3D人体姿态估计

DOPE: Distillation Of Part Experts for whole-body 3D pose estimation in the wild

• 主页：https://europe.naverlabs.com/blog/dope-distillation-of-part-experts-for-whole-body-3d-pose-estimation-in-the-wild/

• 论文：https://arxiv.org/abs/2008.09457

• 代码：https://github.com/naver/dope

SMAP: Single-Shot Multi-Person Absolute 3D Pose Estimation

• 主页：https://zju3dv.github.io/SMAP/
• 论文：https://arxiv.org/abs/2008.11469
• 代码：https://github.com/zju3dv/SMAP

6D位姿估计

CosyPose: Consistent multi-view multi-object 6D pose estimation

• 主页：https://www.di.ens.fr/willow/research/cosypose/

• 论文：http://xxx.itp.ac.cn/abs/2008.08465

• 代码：https://github.com/ylabbe/cosypose

深度估计

Learning Stereo from Single Images

• 论文：https://arxiv.org/abs/2008.01484
• 代码：https://github.com/nianticlabs/stereo-from-mono/

单目深度估计

P2Net: Patch-match and Plane-regularization for Unsupervised Indoor Depth Estimation

• 论文：https://arxiv.org/abs/2007.07696
• 代码：https://github.com/svip-lab/Indoor-SfMLearner

Self-Supervised Monocular Depth Estimation: Solving the Dynamic Object Problem by Semantic Guidance

• 论文：https://arxiv.org/abs/2007.06936

• 代码：https://github.com/ifnspaml/SGDepth

深度补全

Non-Local Spatial Propagation Network for Depth Completion

• 论文：https://arxiv.org/abs/2007.10042
• 代码：https://github.com/zzangjinsun/NLSPN_ECCV20

域泛化

Learning from Extrinsic and Intrinsic Supervisions for Domain Generalization

• 论文：https://arxiv.org/abs/2007.09316
• 代码：https://github.com/EmmaW8/EISNet

超分辨率

图像超分辨率

Learning the Super-Resolution Space with Normalizing Flow

• 论文：https://arxiv.org/abs/2006.14200
• 代码：https://github.com/andreas128/SRFlow

Deep Decomposition Learning for Inverse Imaging Problems

• 论文：https://arxiv.org/abs/1911.11028
• 代码：https://github.com/edongdongchen/DDN

Component Divide-and-Conquer for Real-World Image Super-Resolution

• 论文：https://arxiv.org/abs/2008.01928
• 代码：https://github.com/xiezw5/Component-Divide-and-Conquer-for-Real-World-Image-Super-Resolution
• 数据集：https://github.com/xiezw5/Component-Divide-and-Conquer-for-Real-World-Image-Super-Resolution

Learning with Privileged Information for Efficient Image Super-Resolution

• 主页：https://cvlab.yonsei.ac.kr/projects/PISR/
• 论文：https://arxiv.org/abs/2007.07524
• 代码：https://github.com/cvlab-yonsei/PISR

Spatial-Angular Interaction for Light Field Image Super-Resolution

• 论文：https://arxiv.org/abs/1912.07849
• 代码：https://github.com/YingqianWang/LF-InterNet

Invertible Image Rescaling

• 论文：https://arxiv.org/abs/2005.05650
• 代码：https://github.com/pkuxmq/Invertible-Image-Rescaling

视频超分辨率

Video Super-Resolution with Recurrent Structure-Detail Network

• 论文：https://arxiv.org/abs/2008.00455

• 代码：https://github.com/junpan19/RSDN

去模糊

图像去模糊

End-to-end Interpretable Learning of Non-blind Image Deblurring

• 论文：https://arxiv.org/abs/2007.01769
• 代码：暂无（即将出来）

视频去模糊

Efficient Spatio-Temporal Recurrent Neural Network for Video Deblurring

• 论文：https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/5116_ECCV_2020_paper.php
• 代码：https://github.com/zzh-tech/ESTRNN

去雨

Rethinking Image Deraining via Rain Streaks and Vapors

• 论文：https://arxiv.org/abs/2008.00823
• 代码：https://github.com/yluestc/derain

图像/视频恢复

Learning Enriched Features for Real Image Restoration and Enhancement

• 论文：https://arxiv.org/abs/2003.06792
• 代码：https://github.com/swz30/MIRNet

图像/视频修复(补全)

NAS-DIP: Learning Deep Image Prior with Neural Architecture Search

• 主页：https://yunchunchen.github.io/NAS-DIP/
• 论文：https://arxiv.org/abs/2008.11713
• 代码：https://github.com/YunChunChen/NAS-DIP-pytorch

Learning Joint Spatial-Temporal Transformations for Video Inpainting

• 论文：https://arxiv.org/abs/2007.10247

• 代码：https://github.com/researchmm/STTN

Rethinking Image Inpainting via a Mutual Encoder-Decoder with Feature Equalizations

• Oral
• 论文：暂无
• 代码：https://github.com/KumapowerLIU/ECCV2020oralRethinking-Image-Inpainting-via-a-Mutual-Encoder-Decoder-with-Feature-Equalizations

风格迁移

Domain-Specific Mappings for Generative Adversarial Style Transfer

• 主页：https://acht7111020.github.io/DSMAP-demo/
• 论文：http://xxx.itp.ac.cn/abs/2008.02198
• 代码：https://github.com/acht7111020/DSMAP

三维重建

Atlas: End-to-End 3D Scene Reconstruction from Posed Images

• 主页：http://zak.murez.com/atlas/
• 论文：https://arxiv.org/abs/2003.10432
• 代码：https://github.com/magicleap/Atlas
• 视频：https://youtu.be/9NOPcOGV6nU

3D Bird Reconstruction: a Dataset, Model, and Shape Recovery from a Single View

• 主页：https://marcbadger.github.io/avian-mesh/
• 论文：https://arxiv.org/abs/2008.06133
• 代码：https://github.com/marcbadger/avian-mesh
• 数据集：https://drive.google.com/file/d/1vyXYIJIo9jneIqC7lowB4GVi17rjztjn/view?usp=sharing

Stochastic Bundle Adjustment for Efficient and Scalable 3D Reconstruction

• 论文：http://xxx.itp.ac.cn/abs/2008.00446

• 代码：https://github.com/zlthinker/STBA

图像描述

Fashion Captioning: Towards Generating Accurate Descriptions with Semantic Rewards

• 论文：https://arxiv.org/abs/2008.02693

• 代码： https://github.com/xuewyang/Fashion_Captioning

• 数据集：https://drive.google.com/drive/folders/1J6SZOt_WFwZToX1Jf7QiXzFVwt23lGwW?usp=sharing

图像检索

SOLAR: Second-Order Loss and Attention for Image Retrieval

• 论文：https://arxiv.org/abs/2001.08972

• 代码：https://github.com/tonyngjichun/SOLAR

Self-supervising Fine-grained Region Similarities for Large-scale Image Localization

• 主页：https://yxgeee.github.io/projects/sfrs

• 论文：https://arxiv.org/abs/2006.03926

• 代码：https://github.com/yxgeee/SFRS

光流估计

RAFT: Recurrent All-Pairs Field Transforms for Optical Flow

• 论文：https://arxiv.org/abs/2003.12039

• 代码：https://github.com/princeton-vl/RAFT

LiteFlowNet3: Resolving Correspondence Ambiguity for More Accurate Optical Flow Estimation

• 论文：https://arxiv.org/abs/2007.09319
• 代码：https://github.com/twhui/LiteFlowNet3

视频插帧

BMBC: Bilateral Motion Estimation with Bilateral Cost Volume for Video Interpolation

• 论文：https://arxiv.org/abs/2007.12622

• 代码：https://github.com/JunHeum/BMBC

车道线检测

CurveLane-NAS: Unifying Lane-Sensitive Architecture Search and Adaptive Point Blending

• 论文：https://arxiv.org/abs/2007.12147
• 数据集：https://github.com/xbjxh/curvelanes

Ultra Fast Structure-aware Deep Lane Detection

• 论文：https://arxiv.org/abs/2004.11757

• 代码：https://github.com/cfzd/Ultra-Fast-Lane-Detection

• 论文解读：https://mp.weixin.qq.com/s/TYzDx8R1oUbVr0FxGnFspQ

Gen-LaneNet: a generalized and scalable approach for 3D lane detection

• 论文：https://arxiv.org/abs/2003.10656
• 代码：https://github.com/yuliangguo/Pytorch_Generalized_3D_Lane_Detection
• 数据集：https://github.com/yuliangguo/3D_Lane_Synthetic_Dataset

轨迹预测

SimAug: Learning Robust Representations from 3D Simulation for Pedestrian Trajectory Prediction in Unseen Cameras

• 论文：https://arxiv.org/abs/2004.02022
• 代码：https://github.com/JunweiLiang/Multiverse

线段检测

Deep Hough-Transform Line Priors

• 论文：https://arxiv.org/abs/2007.09493

• 代码：https://github.com/yanconglin/Deep-Hough-Transform-Line-Priors

视线估计

ETH-XGaze: A Large Scale Dataset for Gaze Estimation under Extreme Head Pose and Gaze Variation

• 主页：https://ait.ethz.ch/projects/2020/ETH-XGaze

• 论文：https://arxiv.org/abs/2007.15837

眼动追踪

Towards End-to-end Video-based Eye-Tracking

• 主页：https://ait.ethz.ch/projects/2020/EVE/
• 论文：https://arxiv.org/abs/2007.13120

对抗攻击

Adversarial Ranking Attack and Defense

• 论文：https://arxiv.org/abs/2002.11293
• 代码：https://github.com/cdluminate/advrank

Square Attack: a query-efficient black-box adversarial attack via random search

• 论文：https://arxiv.org/abs/1912.00049
• 代码：https://github.com/max-andr/square-attack

数据集

Long-term Human Motion Prediction with Scene Context

• 主页：https://people.eecs.berkeley.edu/~zhecao/hmp/index.html

• 论文：https://arxiv.org/abs/2007.03672

• 数据集：https://github.com/ZheC/GTA-IM-Dataset

Object Detection with a Unified Label Space from Multiple Datasets

• 主页：http://www.nec-labs.com/~mas/UniDet/
• 论文：https://arxiv.org/abs/2008.06614
• 代码：暂无
• 数据集：http://www.nec-labs.com/~mas/UniDet/resources/UOD_dataset_ECCV20.zip

Simulating Content Consistent Vehicle Datasets with Attribute Descent

• 论文：https://arxiv.org/abs/1912.08855
• 代码：https://github.com/yorkeyao/VehicleX
• 数据集：https://github.com/yorkeyao/VehicleX

InterHand2.6M: A Dataset and Baseline for 3D Interacting Hand Pose Estimation from a Single RGB Image

• 主页：https://mks0601.github.io/InterHand2.6M/
• 论文：https://arxiv.org/abs/2008.09309
• 代码：https://github.com/facebookresearch/InterHand2.6M

SNE-RoadSeg: Incorporating Surface Normal Information into Semantic Segmentation for Accurate Freespace Detection

• 主页：https://sites.google.com/view/sne-roadseg
• 论文：https://arxiv.org/abs/2008.11351
• 代码：https://github.com/hlwang1124/SNE-RoadSeg
• 数据集：https://drive.google.com/file/d/1YnkqPmzxtjNfMi2B1gMy7LQa5Gnu-BsH/view

CurveLane-NAS: Unifying Lane-Sensitive Architecture Search and Adaptive Point Blending

• 论文：https://arxiv.org/abs/2007.12147
• 数据集：https://github.com/xbjxh/curvelanes

Detecting natural disasters, damage, and incidents in the wild

• 主页：http://incidentsdataset.csail.mit.edu/
• 论文：https://arxiv.org/abs/2008.09188
• 数据集：https://github.com/ethanweber/IncidentsDataset

Simultaneous Detection and Tracking with Motion Modelling for Multiple Object Tracking

• 论文：https://arxiv.org/abs/2008.08826
• 代码：https://github.com/shijieS/DMMN
• 数据集：https://github.com/shijieS/OmniMOTDataset

3D Bird Reconstruction: a Dataset, Model, and Shape Recovery from a Single View

• 主页：https://marcbadger.github.io/avian-mesh/
• 论文：https://arxiv.org/abs/2008.06133
• 代码：https://github.com/marcbadger/avian-mesh
• 数据集：https://drive.google.com/file/d/1vyXYIJIo9jneIqC7lowB4GVi17rjztjn/view?usp=sharing

Fashion Captioning: Towards Generating Accurate Descriptions with Semantic Rewards

• 论文：https://arxiv.org/abs/2008.02693

• 代码： https://github.com/xuewyang/Fashion_Captioning

• 数据集：https://drive.google.com/drive/folders/1J6SZOt_WFwZToX1Jf7QiXzFVwt23lGwW?usp=sharing

From Shadow Segmentation to Shadow Removal

• 论文：http://xxx.itp.ac.cn/abs/2008.00267

• 数据集：https://www3.cs.stonybrook.edu/~cvl/projects/FSS2SR/index.html

LEMMA: A Multi-view Dataset for Learning Multi-agent Multi-task Activities

• 主页：https://sites.google.com/view/lemma-activity

• 论文：https://arxiv.org/abs/2007.15781

• 数据集：https://sites.google.com/view/lemma-activity/home/dataset

• 代码：https://github.com/Buzz-Beater/LEMMA

Component Divide-and-Conquer for Real-World Image Super-Resolution

• 论文：https://arxiv.org/abs/2008.01928

• 代码和数据集：https://github.com/xiezw5/Component-Divide-and-Conquer-for-Real-World-Image-Super-Resolution

Towards End-to-end Video-based Eye-Tracking

• 主页：https://ait.ethz.ch/projects/2020/EVE/
• 论文：https://arxiv.org/abs/2007.13120

Reconstructing NBA Players

• 主页：http://grail.cs.washington.edu/projects/nba_players/

• 论文：https://arxiv.org/abs/2007.13303

CelebA-Spoof: Large-Scale Face Anti-Spoofing Dataset with Rich Annotations

• 论文：https://arxiv.org/abs/2007.12342

• 数据集：https://github.com/Davidzhangyuanhan/CelebA-Spoof

PIoU Loss: Towards Accurate Oriented Object Detection in Complex Environments

• 论文：https://arxiv.org/abs/2007.09584

• 代码：https://github.com/clobotics/piou

• 数据集：https://github.com/clobotics/piou

DanbooRegion: An Illustration Region Dataset

• 主页：https://lllyasviel.github.io/DanbooRegion/

• 论文：https://lllyasviel.github.io/DanbooRegion/paper/paper.pdf

• 数据集：https://github.com/lllyasviel/DanbooRegion

Segment as Points for Efficient Online Multi-Object Tracking and Segmentation

• Oral
• 论文：https://arxiv.org/abs/2007.01550
• 代码：https://github.com/detectRecog/PointTrack
• 数据集：https://github.com/detectRecog/PointTrack

Gen-LaneNet: a generalized and scalable approach for 3D lane detection

• 论文：https://arxiv.org/abs/2003.10656
• 代码：https://github.com/yuliangguo/Pytorch_Generalized_3D_Lane_Detection
• 数据集：https://github.com/yuliangguo/3D_Lane_Synthetic_Dataset

TAO: A Large-Scale Benchmark for Tracking Any Object

• 主页：http://taodataset.org/
• 论文：https://arxiv.org/abs/2005.10356
• 代码：https://github.com/TAO-Dataset/tao

Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling

• 主页：http://structured3d-dataset.org
• 论文：https://arxiv.org/abs/1908.00222
• 代码：https://github.com/bertjiazheng/Structured3D

AiR: Attention with Reasoning Capability

• 论文：暂无

• 代码：https://github.com/szzexpoi/AiR

• 数据集：https://github.com/szzexpoi/AiR

其他

Defocus Blur Detection via Depth Distillation

• 论文：https://arxiv.org/abs/2007.08113
• 代码：https://github.com/vinthony/depth-distillation

Pose Augmentation: Class-agnostic Object Pose Transformation for Object Recognition

• 论文：https://arxiv.org/abs/2003.08526

• 代码：https://github.com/gyhandy/Pose-Augmentation

Improving Multispectral Pedestrian Detection by Addressing Modality Imbalance Problems

• 论文：https://arxiv.org/abs/2008.03043
• 代码：https://github.com/CalayZhou/MBNet
• Demo：https://www.bilibili.com/video/BV1Hi4y137aS

From Shadow Segmentation to Shadow Removal

论文：http://xxx.itp.ac.cn/abs/2008.00267

代码和数据集：https://www3.cs.stonybrook.edu/~cvl/projects/FSS2SR/index.html

Funnel Activation for Visual Recognition

• 论文：https://arxiv.org/abs/2007.11824

• 代码：https://github.com/megvii-model/FunnelAct

Open-Edit: Open-Domain Image Manipulation with Open-Vocabulary Instructions

• 论文：暂无
• 代码：https://github.com/xh-liu/Open-Edit
• Video：https://youtu.be/8E3bwvjCHYE

Consensus-Aware Visual-Semantic Embedding for Image-Text Matching

• 论文：https://arxiv.org/abs/2007.08883
• 代码：https://github.com/BruceW91/CVSE

Perceiving 3D Human-Object Spatial Arrangements from a Single Image in the Wild

• 主页：https://jasonyzhang.com/phosa/
• 论文下载：https://arxiv.org/abs/2007.15649
• 代码：https://github.com/jasonyzhang/phosa

AiR: Attention with Reasoning Capability

• Oral

• 论文：https://arxiv.org/abs/2007.14419

• 代码：https://github.com/szzexpoi/AiR

Distribution-Balanced Loss for Multi-Label Classification in Long-Tailed Datasets

• 论文：https://arxiv.org/abs/2007.09654
• 代码：https://github.com/wutong16/DistributionBalancedLoss

A Generic Visualization Approach for Convolutional Neural Networks

• 论文：https://arxiv.org/abs/2007.09748

• 代码：https://github.com/ahmdtaha/constrained_attention_filter

Deep Plastic Surgery: Robust and Controllable Image Editing with Human-Drawn Sketches

• 主页：https://williamyang1991.github.io/projects/ECCV2020
• 论文：https://arxiv.org/abs/2001.02890
• 代码：https://github.com/TAMU-VITA/DeepPS

GIQA: Generated Image Quality Assessment

• 论文：https://arxiv.org/abs/2003.08932
• 代码：https://github.com/cientgu/GIQA

Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling

• 主页：http://structured3d-dataset.org
• 论文：https://arxiv.org/abs/1908.00222
• 代码：https://github.com/bertjiazheng/Structured3D

AiR: Attention with Reasoning Capability

• 论文：暂无

• 代码：https://github.com/szzexpoi/AiR

• 数据集：https://github.com/szzexpoi/AiR

不确定中没中

Relation Aware Panoptic Segmentation

• 论文：暂无
• 代码：https://github.com/RAPNet/RAP

Spatial-Angular Interaction for Light Field Image Super-Resolution

• 论文：暂无
• 代码：https://github.com/YingqianWang/LF-InterNet

TVR: A Large-Scale Dataset for Video-Subtitle Moment Retrieval

• 论文：https://arxiv.org/abs/2001.09099
• 代码：https://github.com/jayleicn/TVRetrieval
• 代码：https://github.com/jayleicn/TVCaption

Self-supervising Fine-grained Region Similarities for IBL

• 论文：暂无
• 代码： https://github.com/ID2191/ECCV2020

https://github.com/lelechen63/eccv2020