CVPR2020在2月24日公佈了所有接受論文ID,相關報道:1470篇!CVPR2020結果出爐,你中了嗎?(附部分論文鏈接/開源代碼/解讀)。自論文ID公佈以來,許多開發者都分享了自己的優秀工作。
從論文ID公佈以來,極市一直在對CVPR進行實時跟進,本文是對80篇CVPR2020論文整理和分類,均有論文鏈接,部分含開源代碼,涵蓋的方向有:目標檢測、目標跟蹤、圖像分割、人臉識別、姿態估計、三維點雲、視頻分析、模型加速、GAN、OCR等方向,還有論文合集打包下載,分享給大家學習。
此外,我們也會在Github和極市社區上保持更新,歡迎大家關注:
https://github.com/extreme-assistant/cvpr2020/blob/master/CVPR2020.md
關注極市平臺,回覆CVPR2020,獲取CVPR2020論文合集下載鏈接。
聲明:本文爲極市平臺原創整理,未經許可,不得擅自轉載。
目錄(點擊標題即可跳轉)
1. 目標檢測
2. 圖像分割
3. 人臉識別
4. 目標跟蹤
5. 三維點雲/三維重建
6. 圖像處理
7. 圖像分類
8. 姿態估計/動作識別
9. 視頻分析
10. OCR
11. GAN
12. 小樣本/零樣本
13. 弱監督/無監督/自監督
14. 行人跟蹤/行人檢測/ReID
15. 神經網絡/模型加速/模型壓縮
16. 超分辨率
17. 視覺常識
1. 目標檢測
-
Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection
論文地址:https://arxiv.org/abs/1912.02424
代碼:https://github.com/sfzhang15/ATSS -
Few-Shot Object Detection with Attention-RPN and Multi-Relation Detector
論文地址:https://arxiv.org/abs/1908.01998
圖像分割
-
Semi-Supervised Semantic Image Segmentation with Self-correcting Networks
論文地址:https://arxiv.org/abs/1811.07073 -
Deep Snake for Real-Time Instance Segmentation
論文地址:https://arxiv.org/abs/2001.01629 -
CenterMask : Real-Time Anchor-Free Instance Segmentation
論文地址:https://arxiv.org/abs/1911.06667
代碼:https://github.com/youngwanLEE/CenterMask -
SketchGCN: Semantic Sketch Segmentation with Graph Convolutional Networks
論文地址:https://arxiv.org/abs/2003.00678 -
PolarMask: Single Shot Instance Segmentation with Polar Representation
論文地址:https://arxiv.org/abs/1909.13226
代碼:https://github.com/xieenze/PolarMask -
xMUDA: Cross-Modal Unsupervised Domain Adaptation for 3D Semantic Segmentation
論文地址:https://arxiv.org/abs/1911.12676 -
BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation
論文地址:https://arxiv.org/abs/2001.00309
人臉識別
-
Towards Universal Representation Learning for Deep Face Recognition
論文地址:https://arxiv.org/abs/2002.11841 -
Suppressing Uncertainties for Large-Scale Facial Expression Recognition
論文地址:https://arxiv.org/abs/2002.10392
代碼:https://github.com/kaiwang960112/Self-Cure-Network -
Face X-ray for More General Face Forgery Detection
論文地址:https://arxiv.org/pdf/1912.13458.pdf
目標跟蹤
- ROAM: Recurrently Optimizing Tracking Model
論文地址:https://arxiv.org/abs/1907.12006
三維點雲&重建
-
PF-Net: Point Fractal Network for 3D Point Cloud Completion
論文地址:https://arxiv.org/abs/2003.00410 -
PointAugment: an Auto-Augmentation Framework for Point Cloud Classification
論文地址:https://arxiv.org/abs/2002.10876
代碼:https://github.com/liruihui/PointAugment/ -
Learning multiview 3D point cloud registration
論文地址:https://arxiv.org/abs/2001.05119 -
C-Flow: Conditional Generative Flow Models for Images and 3D Point Clouds
論文地址:https://arxiv.org/abs/1912.07009 -
RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds
論文地址:https://arxiv.org/abs/1911.11236 -
Total3DUnderstanding: Joint Layout, Object Pose and Mesh Reconstruction for Indoor Scenes from a Single Image
論文地址:https://arxiv.org/abs/2002.12212 -
Implicit Functions in Feature Space for 3D Shape Reconstruction and Completion
論文地址:https://arxiv.org/abs/2003.01456 -
In Perfect Shape: Certifiably Optimal 3D Shape Reconstruction from 2D Landmarks
論文地址:https://arxiv.org/pdf/1911.11924.pdf -
Attentive Context Normalization for Robust Permutation-Equivariant Learning
論文地址:https://arxiv.org/abs/1907.02545 Weiwei Sun, Wei Jiang, Eduard Trulls, Andrea Tagliasacchi, Kwang Moo Yi
10.PQ-NET: A Generative Part Seq2Seq Network for 3D Shapes
https://arxiv.org/abs/1911.10949
圖像處理
-
Learning to Shade Hand-drawn Sketches
論文地址:https://arxiv.org/abs/2002.11812 -
Single Image Reflection Removal through Cascaded Refinement
論文地址:https://arxiv.org/abs/1911.06634 -
Generalized ODIN: Detecting Out-of-distribution Image without Learning from Out-of-distribution Data
論文地址:https://arxiv.org/abs/2002.11297 -
Deep Image Harmonization via Domain Verification
論文地址:https://arxiv.org/abs/1911.13239
代碼:https://github.com/bcmi/Image_Harmonization_Datasets -
RoutedFusion: Learning Real-time Depth Map Fusion
論文地址:https://arxiv.org/pdf/2001.04388.pdf
圖像分類
-
Self-training with Noisy Student improves ImageNet classification
論文地址:https://arxiv.org/abs/1911.04252 -
Image Matching across Wide Baselines: From Paper to Practice
論文地址:https://arxiv.org/abs/2003.01587 -
Towards Robust Image Classification Using Sequential Attention Models
論文地址:https://arxiv.org/abs/1912.02184 -
Learning in the Frequency Domain
論文地址:https://arxiv.org/abs/2002.12416 -
Learning from Web Data with Memory Module
論文地址:https://arxiv.org/abs/1906.12028 -
Making Better Mistakes: Leveraging Class Hierarchies with Deep Networks
論文地址:https://arxiv.org/abs/1912.09393
姿態估計/動作識別
-
VIBE: Video Inference for Human Body Pose and Shape Estimation
論文地址:https://arxiv.org/abs/1912.05656
代碼:https://github.com/mkocabas/VIBE -
Distribution-Aware Coordinate Representation for Human Pose Estimation
論文地址:https://arxiv.org/abs/1910.06278
代碼:https://github.com/ilovepose/DarkPose -
4D Association Graph for Realtime Multi-person Motion Capture Using Multiple Video Cameras
論文地址:https://arxiv.org/abs/2002.12625 -
Optimal least-squares solution to the hand-eye calibration problem
論文地址:https://arxiv.org/abs/2002.10838 -
D3VO: Deep Depth, Deep Pose and Deep Uncertainty for Monocular Visual Odometry
論文地址:https://arxiv.org/abs/2003.01060 -
Multi-Modal Domain Adaptation for Fine-Grained Action Recognition
論文地址:https://arxiv.org/abs/2001.09691 -
Distribution Aware Coordinate Representation for Human Pose Estimation
論文地址:https://arxiv.org/abs/1910.06278 -
The Devil is in the Details: Delving into Unbiased Data Processing for Human Pose Estimation
論文地址:https://arxiv.org/abs/1911.07524 -
PVN3D: A Deep Point-wise 3D Keypoints Voting Network for 6DoF Pose Estimation
論文地址:https://arxiv.org/abs/1911.04231 -
Action Segmentation with Joint Self-Supervised Temporal Domain Adaptation
論文地址:https://arxiv.org/abs/2003.02824
視頻分析
-
Rethinking Zero-shot Video Classification: End-to-end Training for Realistic Applications
論文地址:https://arxiv.org/abs/2003.01455
代碼:https://github.com/bbrattoli/ZeroShotVideoClassification -
Say As You Wish: Fine-grained Control of Image Caption Generation with Abstract Scene Graphs
論文地址:https://arxiv.org/abs/2003.00387 -
Fine-grained Video-Text Retrieval with Hierarchical Graph Reasoning
論文地址:https://arxiv.org/abs/2003.00392 -
Object Relational Graph with Teacher-Recommended Learning for Video Captioning
論文地址:https://arxiv.org/abs/2002.11566 -
Zooming Slow-Mo: Fast and Accurate One-Stage Space-Time Video Super-Resolution
論文地址:https://arxiv.org/abs/2002.11616 -
Blurry Video Frame Interpolation
論文地址:https://arxiv.org/abs/2002.12259 -
Hierarchical Conditional Relation Networks for Video Question Answering
論文地址:https://arxiv.org/abs/2002.10698 -
Action Modifiers:Learning from Adverbs in Instructional Video
論文地址:https://arxiv.org/abs/1912.06617
OCR
- ABCNet: Real-time Scene Text Spotting with Adaptive Bezier-Curve Network
論文地址:https://arxiv.org/abs/2002.10200
代碼:https://github.com/Yuliang-Liu/bezier_curve_text_spotting,https://github.com/aim-uofa/adet
GAN
-
Your Local GAN: Designing Two Dimensional Local Attention Mechanisms for Generative Models
論文地址:https://arxiv.org/abs/1911.12287
代碼:https://github.com/giannisdaras/ylg -
MSG-GAN: Multi-Scale Gradient GAN for Stable Image Synthesis
論文地址:https://arxiv.org/abs/1903.06048 -
Robust Design of Deep Neural Networks against Adversarial Attacks based on Lyapunov Theory
論文地址:https://arxiv.org/abs/1911.04636
小樣本/零樣本
- Improved Few-Shot Visual Classification
論文地址:https://arxiv.org/pdf/1912.03432.pdf
2.Meta-Transfer Learning for Zero-Shot Super-Resolution
論文地址:https://arxiv.org/abs/2002.12213
弱監督/無監督/自監督
-
Rethinking the Route Towards Weakly Supervised Object Localization
論文地址:https://arxiv.org/abs/2002.11359 -
NestedVAE: Isolating Common Factors via Weak Supervision
論文地址:https://arxiv.org/abs/2002.11576 -
Unsupervised Reinforcement Learning of Transferable Meta-Skills for Embodied Navigation
論文地址:https://arxiv.org/abs/1911.07450 -
Disentangling Physical Dynamics from Unknown Factors for Unsupervised Video Prediction
論文地址:https://arxiv.org/abs/2003.01460 -
ClusterFit: Improving Generalization of Visual Representations
論文地址:https://arxiv.org/abs/1912.03330 -
Auto-Encoding Twin-Bottleneck Hashing
論文地址:https://arxiv.org/abs/2002.11930 -
Learning Representations by Predicting Bags of Visual Words
論文地址:https://arxiv.org/abs/2002.12247 -
A Characteristic Function Approach to Deep Implicit Generative Modeling
論文地址:https://arxiv.org/abs/1909.07425
行人跟蹤/行人檢測/ReID
- Cross-modality Person re-identification with Shared-Specific Feature Transfer
論文地址:https://arxiv.org/abs/2002.12489
2.Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction
論文地址:https://arxiv.org/abs/2002.11927
3.The Garden of Forking Paths: Towards Multi-Future Trajectory Prediction
論文地址:https://arxiv.org/abs/1912.06445
神經網絡/模型壓縮/模型加速
-
GhostNet: More Features from Cheap Operations
論文地址:https://arxiv.org/abs/1911.11907
代碼:https://github.com/iamhankai/ghostnet -
Watch your Up-Convolution: CNN Based Generative Deep Neural Networks are Failing to Reproduce Spectral
論文地址:https://arxiv.org/abs/2003.01826 -
GPU-Accelerated Mobile Multi-view Style Transfer
論文地址:https://arxiv.org/abs/2003.00706 -
Bundle Adjustment on a Graph Processor
論文地址:https://arxiv.org/abs/2003.03134
代碼:https://github.com/joeaortiz/gbp -
Watch your Up-Convolution: CNN Based Generative Deep Neural Networks are Failing to Reproduce Spectral
論文地址:https://arxiv.org/abs/2003.01826 -
Holistically-Attracted Wireframe Parsing
論文地址:https://arxiv.org/abs/2003.01663 -
AdderNet: Do We Really Need Multiplications in Deep Learning?
論文地址:https://arxiv.org/abs/1912.13200 -
CARS: Contunuous Evolution for Efficient Neural Architecture Search
論文地址:https://arxiv.org/abs/1909.04977
代碼:https://github.com/huawei-noah/CARS -
Π-nets: Deep Polynomial Neural Networksv
論文地址:https://arxiv.org/abs/2003.03828
超分辨率
- Zooming Slow-Mo: Fast and Accurate One-Stage Space-Time Video Super-Resolution
論文地址:https://arxiv.org/abs/2002.11616
視覺常識
-
Visual Commonsense R-CNN
論文地址:https://arxiv.org/abs/2002.12204
代碼:https://github.com/Wangt-CN/VC-R-CNN -
Scalable Uncertainty for Computer Vision with Functional Variational Inference
論文地址:https://arxiv.org/abs/2003.03396 -
Deep Representation Learning on Long-tailed Data: A Learnable Embedding Augmentation Perspective
論文地址:https://arxiv.org/abs/2002.10826 -
Representations, Metrics and Statistics For Shape Analysis of Elastic Graphs
論文地址:https://arxiv.org/abs/2003.00287 -
Filter Grafting for Deep Neural Networks
論文地址:https://arxiv.org/abs/2001.05868
代碼:https://github.com/fxmeng/filter-grafting.git -
12-in-1: Multi-Task Vision and Language Representation Learning
論文地址:https://arxiv.org/abs/1912.02315 -
Towards Learning a Generic Agent for Vision-and-Language Navigation via Pre-training
論文地址:https://arxiv.org/abs/2002.10638
代碼:https://github.com/weituo12321/PREVALENT -
Unbiased Scene Graph Generation from Biased Training
論文地址:https://arxiv.org/abs/2002.11949