ngnquan/CVPR2020-Code

CVPR 2020 Open source project collection

CVPR2020-Code

CVPR 2020 Paper Open source project

• CNN
• Image Classification
• Object Detection
• 3D Object Detection
• Video Object Detection
• Object Tracking
• Semantic Segmentation
• Instance Segmentation
• Panoptic Segmentation
• Video Object Segmentation
• Superpixel Segmentation
• NAS
• GAN
• Re-ID
• 3D Point cloud (classification/segmentation/registration, etc.)
• Face (recognition/detection/reconstruction, etc.)
• Human Pose Estimation(2D/3D)
• Human Parsing
• Scene Text Detection
• Scene Text Recognition
• Super Resolution
• Model Compression
• Action Recognition
• Crowd Counting
• Depth Estimation
• 6D Target pose estimation
• Hand Pose
• Saliency Object Detection
• Denoising
• Deblurring
• Dehazing
• Feature point detection and description
• Visual Q&A
• VideoQA
• Visual language navigation
• Video compression
• Video interpolation
• Style Transfer
• Lane Detection
• Human-Object Interaction
• Human Trajectory Prediction
• Motion Prediction
• Virtual Try-On
• HDR
• Adversarial
• Semantic Scene Completion
• Dataset
• Others
• Not Sure

CNN

Exploring Self-attention for Image Recognition

• Paper：https://hszhao.github.io/papers/cvpr20_san.pdf

• Code：https://github.com/hszhao/SAN

Improving Convolutional Networks with Self-Calibrated Convolutions

• Homepage：https://mmcheng.net/scconv/

• Paper：http://mftp.mmcheng.net/Papers/20cvprSCNet.pdf

• Code：https://github.com/backseason/SCNet

Rethinking Depthwise Separable Convolutions: How Intra-Kernel Correlations Lead to Improved MobileNets

• Paper：https://arxiv.org/abs/2003.13549
• Code：https://github.com/zeiss-microscopy/BSConv

Image Classification

Compositional Convolutional Neural Networks: A Deep Architecture with Innate Robustness to Partial Occlusion

• Paper：https://arxiv.org/abs/2003.04490

• Code：https://github.com/AdamKortylewski/CompositionalNets

Spatially Attentive Output Layer for Image Classification

• Paper：https://arxiv.org/abs/2004.07570

• Code（Seems to be deleted by the original author）：https://github.com/ildoonet/spatially-attentive-output-layer

Object Detection

Dynamic Refinement Network for Oriented and Densely Packed Object Detection

• Paper：https://arxiv.org/abs/2005.09973

• Code with Dataset：https://github.com/Anymake/DRN_CVPR2020

Scale-Equalizing Pyramid Convolution for Object Detection

Paper：https://arxiv.org/abs/2005.03101

Code：https://github.com/jshilong/SEPC

Revisiting the Sibling Head in Object Detector

• Paper：https://arxiv.org/abs/2003.07540

• Code：https://github.com/Sense-X/TSD

Detection in Crowded Scenes: One Proposal, Multiple Predictions

• Paper：https://arxiv.org/abs/2003.09163
• Code：https://github.com/megvii-model/CrowdDetection

Instance-aware, Context-focused, and Memory-efficient Weakly Supervised Object Detection

• Paper：https://arxiv.org/abs/2004.04725
• Code：https://github.com/NVlabs/wetectron

Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection

• Paper：https://arxiv.org/abs/1912.02424
• Code：https://github.com/sfzhang15/ATSS

BiDet: An Efficient Binarized Object Detector

• Paper：https://arxiv.org/abs/2003.03961
• Code：https://github.com/ZiweiWangTHU/BiDet

Harmonizing Transferability and Discriminability for Adapting Object Detectors

• Paper：https://arxiv.org/abs/2003.06297
• Code：https://github.com/chaoqichen/HTCN

CentripetalNet: Pursuing High-quality Keypoint Pairs for Object Detection

• Paper：https://arxiv.org/abs/2003.09119
• Code：https://github.com/KiveeDong/CentripetalNet

Hit-Detector: Hierarchical Trinity Architecture Search for Object Detection

• Paper：https://arxiv.org/abs/2003.11818
• Code：https://github.com/ggjy/HitDet.pytorch

EfficientDet: Scalable and Efficient Object Detection

• Paper：https://arxiv.org/abs/1911.09070
• Code：https://github.com/google/automl/tree/master/efficientdet

3D Object Detection

Train in Germany, Test in The USA: Making 3D Object Detectors Generalize

• Paper：https://arxiv.org/abs/2005.08139

• Code：https://github.com/cxy1997/3D_adapt_auto_driving

MLCVNet: Multi-Level Context VoteNet for 3D Object Detection

• Paper：https://arxiv.org/abs/2004.05679
• Code：https://github.com/NUAAXQ/MLCVNet

3DSSD: Point-based 3D Single Stage Object Detector

• CVPR 2020 Oral

• Paper：https://arxiv.org/abs/2002.10187

• Code：https://github.com/tomztyang/3DSSD

Disp R-CNN: Stereo 3D Object Detection via Shape Prior Guided Instance Disparity Estimation

• Paper：https://arxiv.org/abs/2004.03572

• Code：https://github.com/zju3dv/disprcn

End-to-End Pseudo-LiDAR for Image-Based 3D Object Detection

• Paper：https://arxiv.org/abs/2004.03080

• Code：https://github.com/mileyan/pseudo-LiDAR_e2e

DSGN: Deep Stereo Geometry Network for 3D Object Detection

• Paper：https://arxiv.org/abs/2001.03398
• Code：https://github.com/chenyilun95/DSGN

LiDAR-based Online 3D Video Object Detection with Graph-based Message Passing and Spatiotemporal Transformer Attention

• Paper：https://arxiv.org/abs/2004.01389
• Code：https://github.com/yinjunbo/3DVID

PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection

• Paper：https://arxiv.org/abs/1912.13192

• Code：https://github.com/sshaoshuai/PV-RCNN

Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud

• Paper：https://arxiv.org/abs/2003.01251
• Code：https://github.com/WeijingShi/Point-GNN

Video Object Detection

Memory Enhanced Global-Local Aggregation for Video Object Detection

Paper：https://arxiv.org/abs/2003.12063

Code：https://github.com/Scalsol/mega.pytorch

Object Tracking

D3S -- A Discriminative Single Shot Segmentation Tracker

• Paper：https://arxiv.org/abs/1911.08862
• Code：https://github.com/alanlukezic/d3s

ROAM: Recurrently Optimizing Tracking Model

• Paper：https://arxiv.org/abs/1907.12006

• Code：https://github.com/skyoung/ROAM

Siam R-CNN: Visual Tracking by Re-Detection

• Homepage：https://www.vision.rwth-aachen.de/page/siamrcnn
• Paper：https://arxiv.org/abs/1911.12836
• Paper2：https://www.vision.rwth-aachen.de/media/papers/192/siamrcnn.pdf
• Code：https://github.com/VisualComputingInstitute/SiamR-CNN

Cooling-Shrinking Attack: Blinding the Tracker with Imperceptible Noises

• Paper：https://arxiv.org/abs/2003.09595
• Code：https://github.com/MasterBin-IIAU/CSA

High-Performance Long-Term Tracking with Meta-Updater

• Paper：https://arxiv.org/abs/2004.00305

• Code：https://github.com/Daikenan/LTMU

AutoTrack: Towards High-Performance Visual Tracking for UAV with Automatic Spatio-Temporal Regularization

• Paper：https://arxiv.org/abs/2003.12949

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

Probabilistic Regression for Visual Tracking

• Paper：https://arxiv.org/abs/2003.12565
• Code：https://github.com/visionml/pytracking

MAST: A Memory-Augmented Self-supervised Tracker

• Paper：https://arxiv.org/abs/2002.07793
• Code：https://github.com/zlai0/MAST

Siamese Box Adaptive Network for Visual Tracking

• Paper：https://arxiv.org/abs/2003.06761

• Code：https://github.com/hqucv/siamban

Semantic Segmentation

Super-BPD: Super Boundary-to-Pixel Direction for Fast Image Segmentation

• Paper：https://arxiv.org/abs/2006.00303

• Code：https://github.com/JianqiangWan/Super-BPD

Single-Stage Semantic Segmentation from Image Labels

• Paper：https://arxiv.org/abs/2005.08104

• Code：https://github.com/visinf/1-stage-wseg

Learning Texture Invariant Representation for Domain Adaptation of Semantic Segmentation

• Paper：https://arxiv.org/abs/2003.00867
• Code：https://github.com/MyeongJin-Kim/Learning-Texture-Invariant-Representation

MSeg: A Composite Dataset for Multi-domain Semantic Segmentation

• Paper：http://vladlen.info/papers/MSeg.pdf
• Code：https://github.com/mseg-dataset/mseg-api

CascadePSP: Toward Class-Agnostic and Very High-Resolution Segmentation via Global and Local Refinement

• Paper：https://arxiv.org/abs/2005.02551
• Code：https://github.com/hkchengrex/CascadePSP

Unsupervised Intra-domain Adaptation for Semantic Segmentation through Self-Supervision

• Oral
• Paper：https://arxiv.org/abs/2004.07703
• Code：https://github.com/feipan664/IntraDA

Self-supervised Equivariant Attention Mechanism for Weakly Supervised Semantic Segmentation

• Paper：https://arxiv.org/abs/2004.04581
• Code：https://github.com/YudeWang/SEAM

Temporally Distributed Networks for Fast Video Segmentation

• Paper：https://arxiv.org/abs/2004.01800

• Code：https://github.com/feinanshan/TDNet

Context Prior for Scene Segmentation

• Paper：https://arxiv.org/abs/2004.01547

• Code：https://git.io/ContextPrior

Strip Pooling: Rethinking Spatial Pooling for Scene Parsing

• Paper：https://arxiv.org/abs/2003.13328

• Code：https://github.com/Andrew-Qibin/SPNet

Cars Can't Fly up in the Sky: Improving Urban-Scene Segmentation via Height-driven Attention Networks

• Paper：https://arxiv.org/abs/2003.05128
• Code：https://github.com/shachoi/HANet

Learning Dynamic Routing for Semantic Segmentation

• Paper：https://arxiv.org/abs/2003.10401

• Code：https://github.com/yanwei-li/DynamicRouting

Instance Segmentation

PolarMask: Single Shot Instance Segmentation with Polar Representation

• Paper：https://arxiv.org/abs/1909.13226
• Code：https://github.com/xieenze/PolarMask
• Blog：https://zhuanlan.zhihu.com/p/84890413

CenterMask : Real-Time Anchor-Free Instance Segmentation

• Paper：https://arxiv.org/abs/1911.06667
• Code：https://github.com/youngwanLEE/CenterMask

BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation

• Paper：https://arxiv.org/abs/2001.00309
• Code：https://github.com/aim-uofa/AdelaiDet

Deep Snake for Real-Time Instance Segmentation

• Paper：https://arxiv.org/abs/2001.01629
• Code：https://github.com/zju3dv/snake

Mask Encoding for Single Shot Instance Segmentation

• Paper：https://arxiv.org/abs/2003.11712

• Code：https://github.com/aim-uofa/AdelaiDet

Panoptic Segmentation

Pixel Consensus Voting for Panoptic Segmentation

• Paper：https://arxiv.org/abs/2004.01849
• Code：还未公布

BANet: Bidirectional Aggregation Network with Occlusion Handling for Panoptic Segmentation

Paper：https://arxiv.org/abs/2003.14031

Code：https://github.com/Mooonside/BANet

Video Objection Segmentation

A Transductive Approach for Video Object Segmentation

• Paper：https://arxiv.org/abs/2004.07193

• Code：https://github.com/microsoft/transductive-vos.pytorch

State-Aware Tracker for Real-Time Video Object Segmentation

• Paper：https://arxiv.org/abs/2003.00482

• Code：https://github.com/MegviiDetection/video_analyst

Learning Fast and Robust Target Models for Video Object Segmentation

• Paper：https://arxiv.org/abs/2003.00908
• Code：https://github.com/andr345/frtm-vos

Learning Video Object Segmentation from Unlabeled Videos

• Paper：https://arxiv.org/abs/2003.05020
• Code：https://github.com/carrierlxk/MuG

Superpixel

Superpixel Segmentation with Fully Convolutional Networks

• Paper：https://arxiv.org/abs/2003.12929
• Code：https://github.com/fuy34/superpixel_fcn

NAS

AOWS: Adaptive and optimal network width search with latency constraints

• Paper：https://arxiv.org/abs/2005.10481
• Code：https://github.com/bermanmaxim/AOWS

Densely Connected Search Space for More Flexible Neural Architecture Search

• Paper：https://arxiv.org/abs/1906.09607

• Code：https://github.com/JaminFong/DenseNAS

MTL-NAS: Task-Agnostic Neural Architecture Search towards General-Purpose Multi-Task Learning

• Paper：https://arxiv.org/abs/2003.14058

• Code：https://github.com/bhpfelix/MTLNAS

FBNetV2: Differentiable Neural Architecture Search for Spatial and Channel Dimensions

• Paper：https://arxiv.org/abs/2004.05565

• Code：https://github.com/facebookresearch/mobile-vision

Neural Architecture Search for Lightweight Non-Local Networks

• Paper：https://arxiv.org/abs/2004.01961
• Code：https://github.com/LiYingwei/AutoNL

Rethinking Performance Estimation in Neural Architecture Search

• Paper：https://arxiv.org/abs/2005.09917
• Code：https://github.com/zhengxiawu/rethinking_performance_estimation_in_NAS
• Blog1：https://www.zhihu.com/question/372070853/answer/1035234510
• Blog2：https://zhuanlan.zhihu.com/p/111167409

CARS: Continuous Evolution for Efficient Neural Architecture Search

• Paper：https://arxiv.org/abs/1909.04977
• Code（Open source soon）：https://github.com/huawei-noah/CARS

GAN

Semantically Mutil-modal Image Synthesis

• Homepage：http://seanseattle.github.io/SMIS
• Paper：https://arxiv.org/abs/2003.12697
• Code：https://github.com/Seanseattle/SMIS

Unpaired Portrait Drawing Generation via Asymmetric Cycle Mapping

• Paper：https://yiranran.github.io/files/CVPR2020_Unpaired%20Portrait%20Drawing%20Generation%20via%20Asymmetric%20Cycle%20Mapping.pdf
• Code：https://github.com/yiranran/Unpaired-Portrait-Drawing

Learning to Cartoonize Using White-box Cartoon Representations

• Paper：https://github.com/SystemErrorWang/White-box-Cartoonization/blob/master/paper/06791.pdf

• Homepage：https://systemerrorwang.github.io/White-box-Cartoonization/

• Code：https://github.com/SystemErrorWang/White-box-Cartoonization

• Blog：https://zhuanlan.zhihu.com/p/117422157

• Demo video：https://www.bilibili.com/video/av56708333

GAN Compression: Efficient Architectures for Interactive Conditional GANs

• Paper：https://arxiv.org/abs/2003.08936

• Code：https://github.com/mit-han-lab/gan-compression

Watch your Up-Convolution: CNN Based Generative Deep Neural Networks are Failing to Reproduce Spectral Distributions

• Paper：https://arxiv.org/abs/2003.01826
• Code：https://github.com/cc-hpc-itwm/UpConv

Re-ID

COCAS: A Large-Scale Clothes Changing Person Dataset for Re-identification

• Paper：https://arxiv.org/abs/2005.07862

• Dataset：No

Transferable, Controllable, and Inconspicuous Adversarial Attacks on Person Re-identification With Deep Mis-Ranking

• Paper：https://arxiv.org/abs/2004.04199

• Code：https://github.com/whj363636/Adversarial-attack-on-Person-ReID-With-Deep-Mis-Ranking

Pose-guided Visible Part Matching for Occluded Person ReID

• Paper：https://arxiv.org/abs/2004.00230
• Code：https://github.com/hh23333/PVPM

Weakly supervised discriminative feature learning with state information for person identification

• Paper：https://arxiv.org/abs/2002.11939
• Code：https://github.com/KovenYu/state-information

3D Point cloud (classification/segmentation/registration, etc.)

3D Point cloud convolution

Global-Local Bidirectional Reasoning for Unsupervised Representation Learning of 3D Point Clouds

• Paper：https://arxiv.org/abs/2003.12971

• Code：https://github.com/raoyongming/PointGLR

Grid-GCN for Fast and Scalable Point Cloud Learning

• Paper：https://arxiv.org/abs/1912.02984

• Code：https://github.com/Xharlie/Grid-GCN

FPConv: Learning Local Flattening for Point Convolution

• Paper：https://arxiv.org/abs/2002.10701
• Code：https://github.com/lyqun/FPConv

3D Point cloud classification

PointAugment: an Auto-Augmentation Framework for Point Cloud Classification

• Paper：https://arxiv.org/abs/2002.10876
• Code（Open source soon）： https://github.com/liruihui/PointAugment/

3D Point cloud semantic segmentation

RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds

• Paper：https://arxiv.org/abs/1911.11236

• Code：https://github.com/QingyongHu/RandLA-Net

• Blog：https://zhuanlan.zhihu.com/p/105433460

Weakly Supervised Semantic Point Cloud Segmentation:Towards 10X Fewer Labels

• Paper：https://arxiv.org/abs/2004.0409

• Code：https://github.com/alex-xun-xu/WeakSupPointCloudSeg

PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation

• Paper：https://arxiv.org/abs/2003.14032
• Code：https://github.com/edwardzhou130/PolarSeg

Learning to Segment 3D Point Clouds in 2D Image Space

• Paper：https://arxiv.org/abs/2003.05593

• Code：https://github.com/WPI-VISLab/Learning-to-Segment-3D-Point-Clouds-in-2D-Image-Space

3D Point cloud instance segmentation

PointGroup: Dual-Set Point Grouping for 3D Instance Segmentation

• Paper：https://arxiv.org/abs/2004.01658
• Code：https://github.com/Jia-Research-Lab/PointGroup

3D Point cloud registration

D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features

• Paper：https://arxiv.org/abs/2003.03164
• Code：https://github.com/XuyangBai/D3Feat

RPM-Net: Robust Point Matching using Learned Features

• Paper：https://arxiv.org/abs/2003.13479
• Code：https://github.com/yewzijian/RPMNet

3D Point Cloud Completion

Cascaded Refinement Network for Point Cloud Completion

• Paper：https://arxiv.org/abs/2004.03327
• Code：https://github.com/xiaogangw/cascaded-point-completion

3D Point cloud target tracking

P2B: Point-to-Box Network for 3D Object Tracking in Point Clouds

• Paper：https://arxiv.org/abs/2005.13888
• Code：https://github.com/HaozheQi/P2B

Human face

Face recognition

CurricularFace: Adaptive Curriculum Learning Loss for Deep Face Recognition

• Paper：https://arxiv.org/abs/2004.00288

• Code：https://github.com/HuangYG123/CurricularFace

Learning Meta Face Recognition in Unseen Domains

• Paper：https://arxiv.org/abs/2003.07733
• Code：https://github.com/cleardusk/MFR
• Blog：https://mp.weixin.qq.com/s/YZoEnjpnlvb90qSI3xdJqQ

Face Detection

Face live detection

Searching Central Difference Convolutional Networks for Face Anti-Spoofing

• Paper：https://arxiv.org/abs/2003.04092

• Code：https://github.com/ZitongYu/CDCN

Facial expression recognition

Suppressing Uncertainties for Large-Scale Facial Expression Recognition

• Paper：https://arxiv.org/abs/2002.10392

• Code（Open source soon）：https://github.com/kaiwang960112/Self-Cure-Network

Face Rotation

Rotate-and-Render: Unsupervised Photorealistic Face Rotation from Single-View Images

• Paper：https://arxiv.org/abs/2003.08124
• Code：https://github.com/Hangz-nju-cuhk/Rotate-and-Render

Face 3D reconstruction

AvatarMe: Realistically Renderable 3D Facial Reconstruction "in-the-wild"

• Paper：https://arxiv.org/abs/2003.13845
• Dataset：https://github.com/lattas/AvatarMe

FaceScape: a Large-scale High Quality 3D Face Dataset and Detailed Riggable 3D Face Prediction

• Paper：https://arxiv.org/abs/2003.13989
• Code：https://github.com/zhuhao-nju/facescape

Human pose estimation (2D/3D)

2D Human pose estimation

HigherHRNet: Scale-Aware Representation Learning for Bottom-Up Human Pose Estimation

• Paper：https://arxiv.org/abs/1908.10357
• Code：https://github.com/HRNet/HigherHRNet-Human-Pose-Estimation

The Devil is in the Details: Delving into Unbiased Data Processing for Human Pose Estimation

• Paper：https://arxiv.org/abs/1911.07524
• Code：https://github.com/HuangJunJie2017/UDP-Pose
• Blog：https://zhuanlan.zhihu.com/p/92525039

Distribution-Aware Coordinate Representation for Human Pose Estimation

• Homepage：https://ilovepose.github.io/coco/

• Paper：https://arxiv.org/abs/1910.06278

• Code：https://github.com/ilovepose/DarkPose

3D Human pose estimation

Fusing Wearable IMUs with Multi-View Images for Human Pose Estimation: A Geometric Approach

• Homepage：https://www.zhe-zhang.com/cvpr2020

• Paper：https://arxiv.org/abs/2003.11163

• Code：https://github.com/CHUNYUWANG/imu-human-pose-pytorch

Bodies at Rest: 3D Human Pose and Shape Estimation from a Pressure Image using Synthetic Data

• Paper：https://arxiv.org/abs/2004.01166

• Code：https://github.com/Healthcare-Robotics/bodies-at-rest

• Dataset：https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/KOA4ML

Self-Supervised 3D Human Pose Estimation via Part Guided Novel Image Synthesis

• Homepage：http://val.cds.iisc.ac.in/pgp-human/
• Paper：https://arxiv.org/abs/2004.04400

Compressed Volumetric Heatmaps for Multi-Person 3D Pose Estimation

• Paper：https://arxiv.org/abs/2004.00329
• Code：https://github.com/fabbrimatteo/LoCO

VIBE: Video Inference for Human Body Pose and Shape Estimation

• Paper：https://arxiv.org/abs/1912.05656
• Code：https://github.com/mkocabas/VIBE

Back to the Future: Joint Aware Temporal Deep Learning 3D Human Pose Estimation

• Paper：https://arxiv.org/abs/2002.11251
• Code：https://github.com/vnmr/JointVideoPose3D

Cross-View Tracking for Multi-Human 3D Pose Estimation at over 100 FPS

• Paper：https://arxiv.org/abs/2003.03972
• Dataset：No

Human body analysis

Correlating Edge, Pose with Parsing

• Paper：https://arxiv.org/abs/2005.01431

• Code：https://github.com/ziwei-zh/CorrPM

Scene text detection

UnrealText: Synthesizing Realistic Scene Text Images from the Unreal World

• Paper：https://arxiv.org/abs/2003.10608
• Code with Dataset：https://github.com/Jyouhou/UnrealText/

ABCNet: Real-time Scene Text Spotting with Adaptive Bezier-Curve Network

• Paper：https://arxiv.org/abs/2002.10200
• Code（Open source soon）：https://github.com/Yuliang-Liu/bezier_curve_text_spotting
• Code（Open source soon）：https://github.com/aim-uofa/adet

Deep Relational Reasoning Graph Network for Arbitrary Shape Text Detection

• Paper：https://arxiv.org/abs/2003.07493

• Code：https://github.com/GXYM/DRRG

Scene text recognition

SEED: Semantics Enhanced Encoder-Decoder Framework for Scene Text Recognition

• Paper：https://arxiv.org/abs/2005.10977
• Code：https://github.com/Pay20Y/SEED

UnrealText: Synthesizing Realistic Scene Text Images from the Unreal World

• Paper：https://arxiv.org/abs/2003.10608
• Code with Dataset：https://github.com/Jyouhou/UnrealText/

ABCNet: Real-time Scene Text Spotting with Adaptive Bezier-Curve Network

• Paper：https://arxiv.org/abs/2002.10200
• Code（Open source soon）：https://github.com/aim-uofa/adet

Learn to Augment: Joint Data Augmentation and Network Optimization for Text Recognition

• Paper：https://arxiv.org/abs/2003.06606

• Code：https://github.com/Canjie-Luo/Text-Image-Augmentation

Super Resolution

Image Super Resolution

Image Super-Resolution with Cross-Scale Non-Local Attention and Exhaustive Self-Exemplars Mining

• Paper：https://arxiv.org/abs/2006.01424
• Code：https://github.com/SHI-Labs/Cross-Scale-Non-Local-Attention

Structure-Preserving Super Resolution with Gradient Guidance

• Paper：https://arxiv.org/abs/2003.13081

• Code：https://github.com/Maclory/SPSR

Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Analysis and a New Strategy

Paper：https://arxiv.org/abs/2004.00448

Code：https://github.com/clovaai/cutblur

Video Super Resolution

Space-Time-Aware Multi-Resolution Video Enhancement

• Homepage：https://alterzero.github.io/projects/STAR.html
• Paper：http://arxiv.org/abs/2003.13170
• Code：https://github.com/alterzero/STARnet

Zooming Slow-Mo: Fast and Accurate One-Stage Space-Time Video Super-Resolution

• Paper：https://arxiv.org/abs/2002.11616
• Code：https://github.com/Mukosame/Zooming-Slow-Mo-CVPR-2020

Model compression/pruning

DMCP: Differentiable Markov Channel Pruning for Neural Networks

• Paper：https://arxiv.org/abs/2005.03354
• Code：https://github.com/zx55/dmcp

Forward and Backward Information Retention for Accurate Binary Neural Networks

• Paper：https://arxiv.org/abs/1909.10788

• Code：https://github.com/htqin/IR-Net

Towards Efficient Model Compression via Learned Global Ranking

• Paper：https://arxiv.org/abs/1904.12368
• Code：https://github.com/cmu-enyac/LeGR

HRank: Filter Pruning using High-Rank Feature Map

• Paper：http://arxiv.org/abs/2002.10179
• Code：https://github.com/lmbxmu/HRank

GAN Compression: Efficient Architectures for Interactive Conditional GANs

• Paper：https://arxiv.org/abs/2003.08936

• Code：https://github.com/mit-han-lab/gan-compression

Group Sparsity: The Hinge Between Filter Pruning and Decomposition for Network Compression

• Paper：https://arxiv.org/abs/2003.08935

• Code：https://github.com/ofsoundof/group_sparsity

Video understanding/behavior recognition

Intra- and Inter-Action Understanding via Temporal Action Parsing

• Paper：https://arxiv.org/abs/2005.10229
• Homepage with Dataset：https://sdolivia.github.io/TAPOS/

3DV: 3D Dynamic Voxel for Action Recognition in Depth Video

• Paper：https://arxiv.org/abs/2005.05501
• Code：https://github.com/3huo/3DV-Action

FineGym: A Hierarchical Video Dataset for Fine-grained Action Understanding

• Homepage：https://sdolivia.github.io/FineGym/
• Paper：https://arxiv.org/abs/2004.06704

TEA: Temporal Excitation and Aggregation for Action Recognition

• Paper：https://arxiv.org/abs/2004.01398

• Code：https://github.com/Phoenix1327/tea-action-recognition

X3D: Expanding Architectures for Efficient Video Recognition

• Paper：https://arxiv.org/abs/2004.04730

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

Temporal Pyramid Network for Action Recognition

• Homepage：https://decisionforce.github.io/TPN

• Paper：https://arxiv.org/abs/2004.03548

• Code：https://github.com/decisionforce/TPN

Skeleton-based motion recognition

Disentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognition

• Paper：https://arxiv.org/abs/2003.14111
• Code：https://github.com/kenziyuliu/ms-g3d

Crowd counting

Depth estimation

Focus on defocus: bridging the synthetic to real domain gap for depth estimation

• Paper：https://arxiv.org/abs/2005.09623
• Code：https://github.com/dvl-tum/defocus-net

Bi3D: Stereo Depth Estimation via Binary Classifications

• Paper：https://arxiv.org/abs/2005.07274

• Code：https://github.com/NVlabs/Bi3D

AANet: Adaptive Aggregation Network for Efficient Stereo Matching

• Paper：https://arxiv.org/abs/2004.09548
• Code：https://github.com/haofeixu/aanet

Towards Better Generalization: Joint Depth-Pose Learning without PoseNet

• Paper：https://github.com/B1ueber2y/TrianFlow

• Code：https://github.com/B1ueber2y/TrianFlow

Monocular depth estimation

On the uncertainty of self-supervised monocular depth estimation

• Paper：https://arxiv.org/abs/2005.06209
• Code：https://github.com/mattpoggi/mono-uncertainty

3D Packing for Self-Supervised Monocular Depth Estimation

• Paper：https://arxiv.org/abs/1905.02693
• Code：https://github.com/TRI-ML/packnet-sfm
• Demo video：https://www.bilibili.com/video/av70562892/

Domain Decluttering: Simplifying Images to Mitigate Synthetic-Real Domain Shift and Improve Depth Estimation

• Paper：https://arxiv.org/abs/2002.12114
• Code：https://github.com/yzhao520/ARC

6D Target pose estimation

MoreFusion: Multi-object Reasoning for 6D Pose Estimation from Volumetric Fusion

• Paper：https://arxiv.org/abs/2004.04336
• Code：https://github.com/wkentaro/morefusion

EPOS: Estimating 6D Pose of Objects with Symmetries

Homepage：http://cmp.felk.cvut.cz/epos

Paper：https://arxiv.org/abs/2004.00605

G2L-Net: Global to Local Network for Real-time 6D Pose Estimation with Embedding Vector Features

• Paper：https://arxiv.org/abs/2003.11089

• Code：https://github.com/DC1991/G2L_Net

Gesture estimation

HOPE-Net: A Graph-based Model for Hand-Object Pose Estimation

• Paper：https://arxiv.org/abs/2004.00060

• Homepage：http://vision.sice.indiana.edu/projects/hopenet

Monocular Real-time Hand Shape and Motion Capture using Multi-modal Data

• Paper：https://arxiv.org/abs/2003.09572

• Code：https://github.com/CalciferZh/minimal-hand

Salient Object Detection

JL-DCF: Joint Learning and Densely-Cooperative Fusion Framework for RGB-D Salient Object Detection

• Paper：https://arxiv.org/abs/2004.08515

• Code：https://github.com/kerenfu/JLDCF/

UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders

• Homepage：http://dpfan.net/d3netbenchmark/

• Paper：https://arxiv.org/abs/2004.05763

• Code：https://github.com/JingZhang617/UCNet

Denoising

A Physics-based Noise Formation Model for Extreme Low-light Raw Denoising

• Paper：https://arxiv.org/abs/2003.12751

• Code：https://github.com/Vandermode/NoiseModel

CycleISP: Real Image Restoration via Improved Data Synthesis

• Paper：https://arxiv.org/abs/2003.07761

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

Deraining

Multi-Scale Progressive Fusion Network for Single Image Deraining

• Paper：https://arxiv.org/abs/2003.10985

• Code：https://github.com/kuihua/MSPFN

Deblurring

Video Deblurring

Cascaded Deep Video Deblurring Using Temporal Sharpness Prior

• Homepage：https://csbhr.github.io/projects/cdvd-tsp/index.html
• Paper：https://arxiv.org/abs/2004.02501
• Code：https://github.com/csbhr/CDVD-TSP

Dehazing

Multi-Scale Boosted Dehazing Network with Dense Feature Fusion

• Paper：https://arxiv.org/abs/2004.13388

• Code：https://github.com/BookerDeWitt/MSBDN-DFF

Feature point detection and description

ASLFeat: Learning Local Features of Accurate Shape and Localization

• Paper：https://arxiv.org/abs/2003.10071

• Code：https://github.com/lzx551402/aslfeat

Visual Q&A (VQA)

VC R-CNN：Visual Commonsense R-CNN

• Paper：https://arxiv.org/abs/2002.12204
• Code：https://github.com/Wangt-CN/VC-R-CNN

VideoQA

Hierarchical Conditional Relation Networks for Video Question Answering

• Paper：https://arxiv.org/abs/2002.10698
• Code：https://github.com/thaolmk54/hcrn-videoqa

Visual language navigation

Towards Learning a Generic Agent for Vision-and-Language Navigation via Pre-training

• Paper：https://arxiv.org/abs/2002.10638
• Code（Open source soon）：https://github.com/weituo12321/PREVALENT

Video compression

Learning for Video Compression with Hierarchical Quality and Recurrent Enhancement

• Paper：https://arxiv.org/abs/2003.01966
• Code：https://github.com/RenYang-home/HLVC

Video interpolation

Space-Time-Aware Multi-Resolution Video Enhancement

• Homepage：https://alterzero.github.io/projects/STAR.html
• Paper：http://arxiv.org/abs/2003.13170
• Code：https://github.com/alterzero/STARnet

Scene-Adaptive Video Frame Interpolation via Meta-Learning

• Paper：https://arxiv.org/abs/2004.00779
• Code：https://github.com/myungsub/meta-interpolation

Softmax Splatting for Video Frame Interpolation

• Homepage：http://sniklaus.com/papers/softsplat
• Paper：https://arxiv.org/abs/2003.05534
• Code：https://github.com/sniklaus/softmax-splatting

Style Transfer

Diversified Arbitrary Style Transfer via Deep Feature Perturbation

• Paper：https://arxiv.org/abs/1909.08223
• Code：https://github.com/EndyWon/Deep-Feature-Perturbation

Collaborative Distillation for Ultra-Resolution Universal Style Transfer

• Paper：https://arxiv.org/abs/2003.08436

• Code：https://github.com/mingsun-tse/collaborative-distillation

Lane Detection

Inter-Region Affinity Distillation for Road Marking Segmentation

• Paper：https://arxiv.org/abs/2004.05304
• Code：https://github.com/cardwing/Codes-for-IntRA-KD

Human-Object Interaction

Detailed 2D-3D Joint Representation for Human-Object Interaction

• Paper：https://arxiv.org/abs/2004.08154

• Code：https://github.com/DirtyHarryLYL/DJ-RN

Cascaded Human-Object Interaction Recognition

• Paper：https://arxiv.org/abs/2003.04262

• Code：https://github.com/tfzhou/C-HOI

VSGNet: Spatial Attention Network for Detecting Human Object Interactions Using Graph Convolutions

• Paper：https://arxiv.org/abs/2003.05541
• Code：https://github.com/ASMIftekhar/VSGNet

Human Trajectory Prediction

Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction

• Paper：https://arxiv.org/abs/2002.11927
• Code：https://github.com/abduallahmohamed/Social-STGCNN

Motion Prediction

Collaborative Motion Prediction via Neural Motion Message Passing

• Paper：https://arxiv.org/abs/2003.06594
• Code：https://github.com/PhyllisH/NMMP

MotionNet: Joint Perception and Motion Prediction for Autonomous Driving Based on Bird's Eye View Maps

• Paper：https://arxiv.org/abs/2003.06754

• Code：https://github.com/pxiangwu/MotionNet

Virtual Try-On

Towards Photo-Realistic Virtual Try-On by Adaptively Generating↔Preserving Image Content

• Paper：https://arxiv.org/abs/2003.05863
• Code：https://github.com/switchablenorms/DeepFashion_Try_On

HDR

Single-Image HDR Reconstruction by Learning to Reverse the Camera Pipeline

• Homepage：https://www.cmlab.csie.ntu.edu.tw/~yulunliu/SingleHDR

• Paper：https://www.cmlab.csie.ntu.edu.tw/~yulunliu/SingleHDR_/00942.pdf

• Code：https://github.com/alex04072000/SingleHDR

Adversarial

Towards Large yet Imperceptible Adversarial Image Perturbations with Perceptual Color Distance

• Paper：https://arxiv.org/abs/1911.02466
• Code：https://github.com/ZhengyuZhao/PerC-Adversarial

Semantic Scene Completion

3D Sketch-aware Semantic Scene Completion via Semi-supervised Structure Prior

• Paper：https://arxiv.org/abs/2003.14052
• Code：https://github.com/charlesCXK/3D-SketchAware-SSC

Dataset

Intra- and Inter-Action Understanding via Temporal Action Parsing

• Paper：https://arxiv.org/abs/2005.10229
• Homepage with Dataset：https://sdolivia.github.io/TAPOS/

Dynamic Refinement Network for Oriented and Densely Packed Object Detection

• Paper：https://arxiv.org/abs/2005.09973

• Code with Dataset：https://github.com/Anymake/DRN_CVPR2020

COCAS: A Large-Scale Clothes Changing Person Dataset for Re-identification

• Paper：https://arxiv.org/abs/2005.07862

• Dataset：No

KeypointNet: A Large-scale 3D Keypoint Dataset Aggregated from Numerous Human Annotations

• Paper：https://arxiv.org/abs/2002.12687

• Dataset：https://github.com/qq456cvb/KeypointNet

MSeg: A Composite Dataset for Multi-domain Semantic Segmentation

• Paper：http://vladlen.info/papers/MSeg.pdf
• Code：https://github.com/mseg-dataset/mseg-api

AvatarMe: Realistically Renderable 3D Facial Reconstruction "in-the-wild"

• Paper：https://arxiv.org/abs/2003.13845
• Dataset：https://github.com/lattas/AvatarMe

Learning to Autofocus

• Paper：https://arxiv.org/abs/2004.12260
• Dataset：No

FaceScape: a Large-scale High Quality 3D Face Dataset and Detailed Riggable 3D Face Prediction

• Paper：https://arxiv.org/abs/2003.13989
• Code：https://github.com/zhuhao-nju/facescape

Bodies at Rest: 3D Human Pose and Shape Estimation from a Pressure Image using Synthetic Data

• Paper：https://arxiv.org/abs/2004.01166

• Code：https://github.com/Healthcare-Robotics/bodies-at-rest

• Dataset：https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/KOA4ML

FineGym: A Hierarchical Video Dataset for Fine-grained Action Understanding

• Homepage：https://sdolivia.github.io/FineGym/
• Paper：https://arxiv.org/abs/2004.06704

A Local-to-Global Approach to Multi-modal Movie Scene Segmentation

• Homepage：https://anyirao.com/projects/SceneSeg.html

• Paper：https://arxiv.org/abs/2004.02678

• Code：https://github.com/AnyiRao/SceneSeg

Deep Homography Estimation for Dynamic Scenes

• Paper：https://arxiv.org/abs/2004.02132

• Dataset：https://github.com/lcmhoang/hmg-dynamics

Assessing Image Quality Issues for Real-World Problems

• Homepage：https://vizwiz.org/tasks-and-datasets/image-quality-issues/
• Paper：https://arxiv.org/abs/2003.12511

UnrealText: Synthesizing Realistic Scene Text Images from the Unreal World

• Paper：https://arxiv.org/abs/2003.10608
• Code with Dataset：https://github.com/Jyouhou/UnrealText/

PANDA: A Gigapixel-level Human-centric Video Dataset

• Paper：https://arxiv.org/abs/2003.04852

• Dataset：http://www.panda-dataset.com/

IntrA: 3D Intracranial Aneurysm Dataset for Deep Learning

• Paper：https://arxiv.org/abs/2003.02920
• Dataset：https://github.com/intra3d2019/IntrA

Cross-View Tracking for Multi-Human 3D Pose Estimation at over 100 FPS

• Paper：https://arxiv.org/abs/2003.03972
• Dataset：No

Others

Differentiable Volumetric Rendering: Learning Implicit 3D Representations without 3D Supervision

• Paper：http://www.cvlibs.net/publications/Niemeyer2020CVPR.pdf

• Code：https://github.com/autonomousvision/differentiable_volumetric_rendering

QEBA: Query-Efficient Boundary-Based Blackbox Attack

• Paper：https://arxiv.org/abs/2005.14137
• Code：https://github.com/AI-secure/QEBA

Equalization Loss for Long-Tailed Object Recognition

• Paper：https://arxiv.org/abs/2003.05176
• Code：https://github.com/tztztztztz/eql.detectron2

Instance-aware Image Colorization

• Homepage：https://ericsujw.github.io/InstColorization/
• Paper：https://arxiv.org/abs/2005.10825
• Code：https://github.com/ericsujw/InstColorization

Contextual Residual Aggregation for Ultra High-Resolution Image Inpainting

• Paper：https://arxiv.org/abs/2005.09704

• Code：https://github.com/Atlas200dk/sample-imageinpainting-HiFill

Where am I looking at? Joint Location and Orientation Estimation by Cross-View Matching

• Paper：https://arxiv.org/abs/2005.03860
• Code：https://github.com/shiyujiao/cross_view_localization_DSM

Epipolar Transformers

• Paper：https://arxiv.org/abs/2005.04551

• Code：https://github.com/yihui-he/epipolar-transformers

Bringing Old Photos Back to Life

• Homepage：http://raywzy.com/Old_Photo/
• Paper：https://arxiv.org/abs/2004.09484

MaskFlownet: Asymmetric Feature Matching with Learnable Occlusion Mask

• Paper：https://arxiv.org/abs/2003.10955

• Code：https://github.com/microsoft/MaskFlownet

Self-Supervised Viewpoint Learning from Image Collections

• Paper：https://arxiv.org/abs/2004.01793
• Paper2：https://research.nvidia.com/sites/default/files/pubs/2020-03_Self-Supervised-Viewpoint-Learning/SSV-CVPR2020.pdf
• Code：https://github.com/NVlabs/SSV

Towards Discriminability and Diversity: Batch Nuclear-norm Maximization under Label Insufficient Situations

• Oral

• Paper：https://arxiv.org/abs/2003.12237

• Code：https://github.com/cuishuhao/BNM

Towards Learning Structure via Consensus for Face Segmentation and Parsing

• Paper：https://arxiv.org/abs/1911.00957
• Code：https://github.com/isi-vista/structure_via_consensus

Plug-and-Play Algorithms for Large-scale Snapshot Compressive Imaging

• Oral

• Paper：https://arxiv.org/abs/2003.13654

• Code：https://github.com/liuyang12/PnP-SCI

Lightweight Photometric Stereo for Facial Details Recovery

• Paper：https://arxiv.org/abs/2003.12307
• Code：https://github.com/Juyong/FacePSNet

Footprints and Free Space from a Single Color Image

• Paper：https://arxiv.org/abs/2004.06376

• Code：https://github.com/nianticlabs/footprints

Self-Supervised Monocular Scene Flow Estimation

• Paper：https://arxiv.org/abs/2004.04143
• Code：https://github.com/visinf/self-mono-sf

Quasi-Newton Solver for Robust Non-Rigid Registration

• Paper：https://arxiv.org/abs/2004.04322
• Code：https://github.com/Juyong/Fast_RNRR

A Local-to-Global Approach to Multi-modal Movie Scene Segmentation

• Homepage：https://anyirao.com/projects/SceneSeg.html

• Paper：https://arxiv.org/abs/2004.02678

• Code：https://github.com/AnyiRao/SceneSeg

DeepFLASH: An Efficient Network for Learning-based Medical Image Registration

• Paper：https://arxiv.org/abs/2004.02097

• Code：https://github.com/jw4hv/deepflash

Self-Supervised Scene De-occlusion

• Homepage：https://xiaohangzhan.github.io/projects/deocclusion/
• Paper：https://arxiv.org/abs/2004.02788
• Code：https://github.com/XiaohangZhan/deocclusion

Polarized Reflection Removal with Perfect Alignment in the Wild

• Homepage：https://leichenyang.weebly.com/project-polarized.html
• Code：https://github.com/ChenyangLEI/CVPR2020-Polarized-Reflection-Removal-with-Perfect-Alignment

Background Matting: The World is Your Green Screen

• Paper：https://arxiv.org/abs/2004.00626
• Code：http://github.com/senguptaumd/Background-Matting

What Deep CNNs Benefit from Global Covariance Pooling: An Optimization Perspective

• Paper：https://arxiv.org/abs/2003.11241

• Code：https://github.com/ZhangLi-CS/GCP_Optimization

Look-into-Object: Self-supervised Structure Modeling for Object Recognition

• Paper：No
• Code：https://github.com/JDAI-CV/LIO

Video Object Grounding using Semantic Roles in Language Description

• Paper：https://arxiv.org/abs/2003.10606
• Code：https://github.com/TheShadow29/vognet-pytorch

Dynamic Hierarchical Mimicking Towards Consistent Optimization Objectives

• Paper：https://arxiv.org/abs/2003.10739
• Code：https://github.com/d-li14/DHM

SDFDiff: Differentiable Rendering of Signed Distance Fields for 3D Shape Optimization

• Paper：http://www.cs.umd.edu/~yuejiang/papers/SDFDiff.pdf
• Code：https://github.com/YueJiang-nj/CVPR2020-SDFDiff

On Translation Invariance in CNNs: Convolutional Layers can Exploit Absolute Spatial Location

• Paper：https://arxiv.org/abs/2003.07064

• Code：https://github.com/oskyhn/CNNs-Without-Borders

GhostNet: More Features from Cheap Operations

• Paper：https://arxiv.org/abs/1911.11907

• Code：https://github.com/iamhankai/ghostnet

AdderNet: Do We Really Need Multiplications in Deep Learning?

• Paper：https://arxiv.org/abs/1912.13200
• Code：https://github.com/huawei-noah/AdderNet

Deep Image Harmonization via Domain Verification

• Paper：https://arxiv.org/abs/1911.13239
• Code：https://github.com/bcmi/Image_Harmonization_Datasets

Blurry Video Frame Interpolation

• Paper：https://arxiv.org/abs/2002.12259
• Code：https://github.com/laomao0/BIN

Extremely Dense Point Correspondences using a Learned Feature Descriptor

• Paper：https://arxiv.org/abs/2003.00619
• Code：https://github.com/lppllppl920/DenseDescriptorLearning-Pytorch

Filter Grafting for Deep Neural Networks

• Paper：https://arxiv.org/abs/2001.05868
• Code：https://github.com/fxmeng/filter-grafting
• PaperBlog：https://www.zhihu.com/question/372070853/answer/1041569335

Action Segmentation with Joint Self-Supervised Temporal Domain Adaptation

• Paper：https://arxiv.org/abs/2003.02824
• Code：https://github.com/cmhungsteve/SSTDA

Detecting Attended Visual Targets in Video

• Paper：https://arxiv.org/abs/2003.02501

• Code：https://github.com/ejcgt/attention-target-detection

Deep Image Spatial Transformation for Person Image Generation

• Paper：https://arxiv.org/abs/2003.00696
• Code：https://github.com/RenYurui/Global-Flow-Local-Attention

Rethinking Zero-shot Video Classification: End-to-end Training for Realistic Applications

• Paper：https://arxiv.org/abs/2003.01455
• Code：https://github.com/bbrattoli/ZeroShotVideoClassification

https://github.com/charlesCXK/3D-SketchAware-SSC

https://github.com/Anonymous20192020/Anonymous_CVPR5767

https://github.com/avirambh/ScopeFlow

https://github.com/csbhr/CDVD-TSP

https://github.com/ymcidence/TBH

https://github.com/yaoyao-liu/mnemonics

https://github.com/meder411/Tangent-Images

https://github.com/KaihuaTang/Scene-Graph-Benchmark.pytorch

https://github.com/sjmoran/deep_local_parametric_filters

https://github.com/charlesCXK/3D-SketchAware-SSC

https://github.com/bermanmaxim/AOWS

https://github.com/dc3ea9f/look-into-object

Not Sure

FADNet: A Fast and Accurate Network for Disparity Estimation

• Paper：还没出来
• Code：https://github.com/HKBU-HPML/FADNet

https://github.com/rFID-submit/RandomFID：Not Sure

https://github.com/JackSyu/AE-MSR：Not Sure

https://github.com/fastconvnets/cvpr2020：Not Sure

https://github.com/aimagelab/meshed-memory-transformer：Not Sure

https://github.com/TWSFar/CRGNet：Not Sure

https://github.com/CVPR-2020/CDARTS：Not Sure

https://github.com/anucvml/ddn-cvprw2020：Not Sure

https://github.com/dl-model-recommend/model-trust：Not Sure

https://github.com/apratimbhattacharyya18/CVPR-2020-Corr-Prior：Not Sure

https://github.com/onetcvpr/O-Net：Not Sure

https://github.com/502463708/Microcalcification_Detection：Not Sure

https://github.com/anonymous-for-review/cvpr-2020-deep-smoke-machine：Not Sure

https://github.com/anonymous-for-review/cvpr-2020-smoke-recognition-dataset：Not Sure

https://github.com/cvpr-nonrigid/dataset：Not Sure

https://github.com/theFool32/PPBA：Not Sure

https://github.com/Realtime-Action-Recognition/Realtime-Action-Recognition