ICCV2021最新信息及已接收论文/代码

官网链接：http://iccv2021.thecvf.com/home
开会时间：2021年10月11日至17日

:exclamation::exclamation::exclamation::star2::star2::star2:📗📗📗ICCV 2021收录论文已全部公布，下载可在【我爱计算机视觉】后台回复“paper”，即可收到。共计 1612 篇。

:exclamation::exclamation::exclamation::star2::star2::star2:全部论文已粗略分类完毕，请查阅

历年综述论文分类汇总戳这里↘️CV-Surveys施工中~~~~~~~~~~

2022 年论文分类汇总戳这里

↘️CVPR-2022-Papers ↘️WACV-2022-Papers

2021年论文分类汇总戳这里

↘️ICCV-2021-Papers ↘️CVPR-2021-Papers

2020 年论文分类汇总戳这里

↘️CVPR-2020-Papers ↘️ECCV-2020-Papers

目录

65.Optical Flow Estimation(光流估计)

• Separable Flow: Learning Motion Cost Volumes for Optical Flow Estimation
  :star:code
• High-Resolution Optical Flow from 1D Attention and Correlation
  :open_mouth:oral:star:code
• GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning
  :star:code
• Sensor-Guided Optical Flow
  :star:code

64.Anomaly Detection(异常检测)

• 表面异常检测
  • DRÆM – A discriminatively trained reconstru
    :star:code
• 异常检测
  • Weakly Supervised Temporal Anomaly Segmentation with Dynamic Time Warping
  • Learning Unsupervised Metaformer for Anomaly Detection
    解决图像异常的分类或定位

63.Data Augmentation(数据增强)

• DivAug: Plug-In Automated Data Augmentation With Explicit Diversity Maximization
  :star:code
• TrivialAugment: Tuning-Free Yet State-of-the-Art Data Augmentation
  :open_mouth:oral:star:code
• Semantic Aware Data Augmentation for Cell Nuclei Microscopical Images With Artificial Neural Networks
• A Simple Baseline for Semi-Supervised Semantic Segmentation With Strong Data Augmentation

62.Open-Set Recognition(开放集识别)

• OpenGAN: Open-Set Recognition via Open Data Generation
  :trophy:Best Paper Honorable Mention
• Conditional Variational Capsule Network for Open Set Recognition
  :star:code

61.Metric Learning(元学习)

• Do Different Deep Metric Learning Losses Lead to Similar Learned Features?
  :star:code
• Learning With Memory-Based Virtual Classes for Deep Metric Learning
  :star:code

60.Federated Learning(联合学习)

• Federated Learning for Non-IID Data via Unified Feature Learning and Optimization Objective Alignment
• Ensemble Attention Distillation for Privacy-Preserving Federated Learning

59.Graph Neural Networks(图神经网络)

• Meta-Aggregator: Learning to Aggregate for 1-bit Graph Neural Networks
• PoGO-Net: Pose Graph Optimization With Graph Neural Networks
  :star:code
• Dynamic Dual Gating Neural Networks
  :star:code

58.Computational Photography(光学、几何、光场成像、计算摄影)

• An Asynchronous Kalman Filter for Hybrid Event Cameras
  :star:code
• 4D Cloud Scattering Tomography
• Snapshot compressive imaging(快照压缩成像)
  • Dense Deep Unfolding Network with 3D-CNN Prior for Snapshot Compressive Imaging
    :star:code
• 光场
  • Light Field Saliency Detection with Dual Local Graph Learning andReciprocative Guidance
  • Fast Light-Field Disparity Estimation With Multi-Disparity-Scale Cost Aggregation
    :star:code
  • SeLFVi: Self-supervised Light-Field Video Reconstruction from Stereo Video
  • SIGNET: Efficient Neural Representation for Light Fields
  • 光场重建
    • Learning Dynamic Interpolation for Extremely Sparse Light Fields With Wide Baselines
      :star:code
• 压缩成像
  • Time-Multiplexed Coded Aperture Imaging: Learned Coded Aperture and Pixel Exposures for Compressive Imaging Systems
• Homography Estimation
  • LocalTrans: A Multiscale Local Transformer Network for Cross-Resolution Homography Estimation
• 计算成像
  • Extreme-Quality Computational Imaging via Degradation Framework
• 光学像差矫正
  • Universal and Flexible Optical Aberration Correction Using Deep-Prior Based Deconvolution
    :star:code

57.Image Matching(图像匹配)

• Matching in the Dark: A Dataset for Matching Image Pairs of Low-light Scenes
• 特征点匹配
  • P2-Net: Joint Description and Detection of Local Features for Pixel and Point Matching
    :star:code

56.Dataset(数据集)

• Large Scale Multi-Illuminant (LSMI) Dataset for Developing White Balance Algorithm Under Mixed Illumination
  :sunflower:dataset
• FloW: A Dataset and Benchmark for Floating Waste Detection in Inland Waters
  :sunflower:dataset
  内陆水域漂浮废物检测数据集和基准
• FloorPlanCAD: A Large-Scale CAD Drawing Dataset for Panoptic Symbol Spotting
  :house:project
• 生物医学图像
  • BioFors: A Large Biomedical Image Forensics Dataset
• 3D重建
  • Common Objects in 3D: Large-Scale Learning and Evaluation of Real-life 3D Category Reconstruction
    :sunflower:dataset
• 航空影像数据集
  • Beyond Road Extraction: A Dataset for Map Update using Aerial Images
    :star:code:house:project
    用于使用航拍图像更新地图的数据集
• 动作识别
  • HAA500: Human-Centric Atomic Action Dataset with Curated Videos
• 目标识别
  • ORBIT: A Real-World Few-Shot Dataset for Teachable Object Recognition
    :star:code:sunflower:dataset
• 车道线检测
  • VIL-100: A New Dataset and a Baseline Model for Video Instance Lane Detection
    :sunflower:dataset
• 自动驾驶
  • Large Scale Interactive Motion Forecasting for Autonomous Driving: The Waymo Open Motion Dataset
• 视觉语言数据集
  • E-ViL: A Dataset and Benchmark for Natural Language Explanations in Vision-Language Tasks
    :star:codeVL
• DeepFake检测
  • KoDF: A Large-Scale Korean DeepFake Detection Dataset
    :sunflower:dataset
• 高质量视频
  • Seeing Dynamic Scene in the Dark: A High-Quality Video Dataset With Mechatronic Alignment
    :sunflower:dataset视频

55.Activity Recognition(活动识别)

• Selective Feature Compression for Efficient Activity Recognition Inference
• 小组活动识别
  • Spatio-Temporal Dynamic Inference Network for Group Activity Recognition
    :star:code
  • GroupFormer: Group Activity Recognition with Clustered Spatial-Temporal Transformer
    :star:code

54.Sketch recognition(草图)

• SketchLattice: Latticed Representation for Sketch Manipulation
• SketchAA: Abstract Representation for Abstract Sketches

53.Vision Localization(视觉定位)

• Continual Learning for Image-Based Camera Localization
  :star:code
• CrowdDriven: A New Challenging Dataset for Outdoor Visual Localization
  :sunflower:dataset
• Pose Correction for Highly Accurate Visual Localization in Large-Scale Indoor Spaces
  :star:code
• Cross-Descriptor Visual Localization and Mapping

52.Vision-and-Language(视觉语言)

• YouRefIt: Embodied Reference Understanding with Language and Gesture
  :open_mouth:oral:house:project
• VLGrammar: Grounded Grammar Induction of Vision and Language
  :star:code
• COOKIE: Contrastive Cross-Modal Knowledge Sharing Pre-Training for Vision-Language Representation
  :star:code
• Panoptic Narrative Grounding
  :open_mouth:oral:star:code
• AESOP: Abstract Encoding of Stories, Objects, and Pictures
  :star:code:tv:video
• Adaptive Hierarchical Graph Reasoning With Semantic Coherence for Video-and-Language Inference
• 视觉推理
  • Interpretable Visual Reasoning via Induced Symbolic Space
• 语义导航
  • THDA: Treasure Hunt Data Augmentation for Semantic Navigation
• 视觉语言导航
  • Airbert: In-domain Pretraining for Vision-and-Language Navigation
    :house:project
  • Waypoint Models for Instruction-guided Navigation in Continuous Environments
    :open_mouth:oral:star:code:house:project:tv:video
  • The Road To Know-Where: An Object-and-Room Informed Sequential BERT for Indoor Vision-Language Navigation
    :star:code
  • Vision-Language Navigation With Random Environmental Mixup
• 视觉对话导航
  • Self-Motivated Communication Agent for Real-World Vision-Dialog Navigation
• 视觉导航
  • Pose Invariant Topological Memory for Visual Navigation
    :star:code
  • Visual Graph Memory With Unsupervised Representation for Visual Navigation
    :star:code:house:project:tv:video
• visual grounding
  • InstanceRefer: Cooperative Holistic Understanding for Visual Grounding on Point Clouds through Instance Multi-level Contextual Referring
    :star:code
  • TransVG: End-to-End Visual Grounding With Transformers
    :star:code
• 视觉对话
  • Unified Questioner Transformer for Descriptive Question Generation in Goal-Oriented Visual Dialogue

51.View Synthesis(视图合成)

• Out-of-boundary View Synthesis Towards Full-Frame Video Stabilization
  :star:code
• Deep 3D Mask Volume for View Synthesis of Dynamic Scenes
  :house:project
• Embedding Novel Views in a Single JPEG Image
• Video Autoencoder: self-supervised disentanglement of static 3D structure and motion
  :open_mouth:oral:star:code:house:project:tv:video
• Geometry-Free View Synthesis: Transformers and No 3D Priors
  :star:code
• Dynamic View Synthesis From Dynamic Monocular Video
  :house:project:tv:video
• Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis
  :house:project:tv:video
• Infinite Nature: Perpetual View Generation of Natural Scenes from a Single Image
  :open_mouth:oral:star:code:house:project:tv:video
• Worldsheet: Wrapping the World in a 3D Sheet for View Synthesis from a Single Image
  :open_mouth:oral:star:code:house:project:tv:video

50.Continual Learning(持续学习)

• Online Continual Learning with Natural Distribution Shifts: An Empirical Study with Visual Data
  :star:code
• Continual Learning on Noisy Data Streams via Self-Purified Replay
  :star:code
• Rehearsal Revealed: The Limits and Merits of Revisiting Samples in Continual Learning
  :star:code
• Co2L: Contrastive Continual Learning
  :star:code

49.Human-Object Interaction(人物交互)

• Exploiting Scene Graphs for Human-Object Interaction Detection
  :star:code
• Spatially Conditioned Graphs for Detecting Human-Object Interactions
  :star:code:tv:video
• Virtual Multi-Modality Self-Supervised Foreground Matting for Human-Object Interaction
• Detecting Human-Object Relationships in Videos
• Weakly Supervised Human-Object Interaction Detection in Video via Contrastive Spatiotemporal Regions
  :star:code:house:project:sunflower:dataset
• Discovering Human Interactions With Large-Vocabulary Objects via Query and Multi-Scale Detection
  :star:code
• Visual Relationship Detection Using Part-and-Sum Transformers With Composite QueriesVRD和HOI
• Interaction Compass: Multi-Label Zero-Shot Learning of Human-Object Interactions via Spatial Relations
  :star:code
• H2O
  • H2O: A Benchmark for Visual Human-Human Object Handover Analysis
• Human Interaction Understanding
  • Consistency-Aware Graph Network for Human Interaction Understanding
    :star:code
  • H2O: Two Hands Manipulating Objects for First Person Interaction Recognition
    :house:project
• 手物交互
  • Toward Human-Like Grasp: Dexterous Grasping via Semantic Representation of Object-Hand
  • Reconstructing Hand-Object Interactions in the Wild
    :house:project
  • CPF: Learning a Contact Potential Field To Model the Hand-Object Interaction
    :star:code手物交互
• HOI(行为理解)
  • GeomNet: A Neural Network Based on Riemannian Geometries of SPD Matrix Space and Cholesky Space for 3D Skeleton-Based Interaction Recognition

48.6DoF

• SO-Pose: Exploiting Self-Occlusion for Direct 6D Pose Estimation
  :star:code
• StereOBJ-1M: Large-scale Stereo Image Dataset for 6D Object Pose Estimation
• SGPA: Structure-Guided Prior Adaptation for Category-Level 6D Object Pose Estimation
• RePOSE: Fast 6D Object Pose Refinement via Deep Texture Rendering
  :star:code
• DualPoseNet: Category-Level 6D Object Pose and Size Estimation Using Dual Pose Network With Refined Learning of Pose Consistency
  :star:code
• PR-GCN: A Deep Graph Convolutional Network With Point Refinement for 6D Pose Estimation
• 物体姿势估计
  • CAPTRA: CAtegory-Level Pose Tracking for Rigid and Articulated Objects From Point Clouds
    :open_mouth:oral:star:code:house:project:tv:video

47.NAS

• BN-NAS: Neural Architecture Search with Batch Normalization
  :star:code
• RANK-NOSH: Efficient Predictor-Based Architecture Search via Non-Uniform Successive Halving
• Pi-NAS: Improving Neural Architecture Search by Reducing Supernet Training Consistency Shift
  :star:code
• Evolving Search Space for Neural Architecture Search
  :star:code:tv:video
• FairNAS: Rethinking Evaluation Fairness of Weight Sharing Neural Architecture Search
  :star:code
• GLiT: Neural Architecture Search for Global and Local Image Transformer
  :star:code
• Neural Architecture Search for Joint Human Parsing and Pose Estimation
  :star:code
• Distilling Optimal Neural Networks: Rapid Search in Diverse Spaces
• Learning Latent Architectural Distribution in Differentiable Neural Architecture Search via Variational Information Maximization
• Not All Operations Contribute Equally: Hierarchical Operation-Adaptive Predictor for Neural Architecture Search
• Zen-NAS: A Zero-Shot NAS for High-Performance Image Recognition
  :star:code
• BossNAS: Exploring Hybrid CNN-Transformers With Block-Wisely Self-Supervised Neural Architecture Search
  :star:code
• NAS-OoD: Neural Architecture Search for Out-of-Distribution Generalization
• AutoSpace: Neural Architecture Search With Less Human Interference
  :star:code
• IDARTS: Interactive Differentiable Architecture Search

46.Defect Detection(缺陷检测)

• DRÆM -- A discriminatively trained reconstruction embedding for surface anomaly detection

45.Image Caption(图像字幕)

• Who's Waldo? Linking People Across Text and Images
  :open_mouth:oral:house:project
  :newspaper:解读:ICCV2021 Oral-新任务！新数据集！康奈尔大学提出了类似VG但又不是VG的PVG任务
• Partial Off-Policy Learning: Balance Accuracy and Diversity for Human-Oriented Image Captioning
• Topic Scene Graph Generation by Attention Distillation From Caption
  :star:code
• Understanding and Evaluating Racial Biases in Image Captioning
  :star:code:house:project
• In Defense of Scene Graphs for Image Captioning
  :star:code
• art description generation(艺术描述生成)
  • Explain Me the Painting: Multi-Topic Knowledgeable Art Description Generation
    :star:code
• Change Captioning
  • Viewpoint-Agnostic Change Captioning With Cycle Consistency

44.Human motion prediction(人体运动预测)

• MSR-GCN: Multi-Scale Residual Graph Convolution Networks for Human Motion Prediction
  :star:code
• Stochastic Scene-Aware Motion Prediction
  :star:code:house:project
• Generating Smooth Pose Sequences for Diverse Human Motion Prediction
  :open_mouth:oral:star:code
• TRiPOD: Human Trajectory and Pose Dynamics Forecasting in the Wild
  :house:project
• Motion Prediction using Trajectory Cues
• 3D人体运动预测
  • Contextually Plausible and Diverse 3D Human Motion Prediction

43.Dense Prediction(密集预测)

• FaPN: Feature-aligned Pyramid Network for Dense Image Prediction
  :star:code
• 多任务密集预测
  • Exploring Relational Context for Multi-Task Dense Prediction

42.Representations Learning(表征学习)

• Learning From Noisy Data With Robust Representation Learning
  :star:code
• Self-Supervised Representation Learning From Flow Equivariance
• Exploring Visual Engagement Signals for Representation Learning
  :star:code
• Switchable K-class Hyperplanes for Noise-Robust Representation Learning
  :star:code
• Region Similarity Representation Learning
  :star:code
• Curious Representation Learning for Embodied Intelligence
  :star:code:house:project
• 视觉表征学习
  • Self-Supervised Visual Representations Learning by Contrastive Mask Prediction
    :newspaper:解读:ICCV2021 比MoCo更通用的对比学习范式，中科大&MSRA提出对比学习新方法MaskCo
  • Temporal Knowledge Consistency for Unsupervised Visual Representation Learning
  • Contrasting Contrastive Self-Supervised Representation Learning Pipelines
    :star:code
  • Concept Generalization in Visual Representation Learning
    :house:project
  • Collaborative Unsupervised Visual Representation Learning from Decentralized Data
  • Episodic Transformer for Vision-and-Language Navigation
    :star:code
  • Multi-VAE: Learning Disentangled View-Common and View-Peculiar Visual Representations for Multi-View Clustering
• 视频表示学习
  • Composable Augmentation Encoding for Video Representation Learning
  • Motion-Focused Contrastive Learning of Video Representations
• ASCNet: Self-Supervised Video Representation Learning With Appearance-Speed Consistency
• ACAV100M: Automatic Curation of Large-Scale Datasets for Audio-Visual Video Representation Learning
  :house:project
• Time-Equivariant Contrastive Video Representation Learning
• Space-Time Crop & Attend: Improving Cross-Modal Video Representation Learning
  :star:code

41.Out-of-Distribution Detection(OOD)

• CODEs: Chamfer Out-of-Distribution Examples against Overconfidence Issue
• Semantically Coherent Out-of-Distribution Detection
  :star:code:house:project
• The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization
  :star:code

40.Metric Learning(度量学习)

• Towards Interpretable Deep Metric Learning with Structural Matching
  :star:code
• Deep Relational Metric Learning
  :star:code
• LoOp: Looking for Optimal Hard Negative Embeddings for Deep Metric Learning
  :star:code
• Manifold Matching via Deep Metric Learning for Generative Modeling
  :star:code

39.Incremental Learning(增量学习)

• 类增量学习
  • Always Be Dreaming: A New Approach for Data-Free Class-Incremental Learning
    :newspaper:解读:让模型实现“终生学习”，佐治亚理工学院提出Data-Free的增量学习
  • Striking a Balance Between Stability and Plasticity for Class-Incremental Learning
  • Synthesized Feature Based Few-Shot Class-Incremental Learning on a Mixture of Subspaces
    :star:code

38.Weakly/Semi-Supervised/Self-supervised/Unsupervised Learning(自/半/弱监督学习)

• 半监督
  • Trash to Treasure: Harvesting OOD Data with Cross-Modal Matching for Open-Set Semi-Supervised Learning
  • Semi-Supervised Learning of Visual Features by Non-Parametrically Predicting View Assignments With Support Samples
    :star:code
  • Semi-Supervised Active Learning for Semi-Supervised Models: Exploit Adversarial Examples With Graph-Based Virtual Labels
  • CoMatch: Semi-Supervised Learning With Contrastive Graph Regularization
    :star:code
  • Multiview Pseudo-Labeling for Semi-supervised Learning from Video
• 自监督
  • Focus on the Positives: Self-Supervised Learning for Biodiversity Monitoring
    :star:code
  • Self-supervised Neural Networks for Spectral Snapshot Compressive Imaging
    :star:code
  • ISD: Self-Supervised Learning by Iterative Similarity Distillation
    :star:code
  • Contrast and Order Representations for Video Self-Supervised Learning
  • On Feature Decorrelation in Self-Supervised Learning
    :open_mouth:oral
  • Geography-Aware Self-Supervised Learning
  • Multimodal Clustering Networks for Self-supervised Learning from Unlabeled Videos
  • Efficient Visual Pretraining with Contrastive Detection
  • Broaden Your Views for Self-Supervised Video Learning
  • CDS: Cross-Domain Self-supervised Pre-training
  • On Compositions of Transformations in Contrastive Self-Supervised Learning
    :star:code
  • Solving Inefficiency of Self-Supervised Representation Learning
    :star:code
  • Divide and Contrast: Self-supervised Learning from Uncurated Data
  • Emerging Properties in Self-Supervised Vision Transformers
    :star:code
  • Mean Shift for Self-Supervised Learning
    :star:code
• 弱监督
  • Weakly Supervised Representation Learning With Coarse Labels
    :star:code

37.Multitask Learning(多任务学习)

• MultiTask-CenterNet (MCN): Efficient and Diverse Multitask Learning using an Anchor Free Approach
  :newspaper:解读:ICCV2021《MultiTask CenterNet》CV多任务新进展！一节更比三节强
• Multi-Task Self-Training for Learning General Representations
  :newspaper:解读:ICCV2021 MuST：还在特定任务里为刷点而苦苦挣扎？谷歌的大佬们都已经开始玩多任务训练了
• UniT: Multimodal Multitask Learning With a Unified Transformer
  :star:code
• Learning Multiple Pixelwise Tasks Based on Loss Scale Balancing
  :star:code
• Learning With Privileged Tasks
• Task Switching Network for Multi-Task Learning

36.SLAM/AR/VR/机器人

• 机器人
  • 室内导航
    • The Surprising Effectiveness of Visual Odometry Techniques for Embodied PointGoal Navigation
      :star:code:house:project
    • Pathdreamer: A World Model for Indoor Navigation
      :tv:video
  • 机器手抓取
    • Hand-Object Contact Consistency Reasoning for Human Grasps Generation
      :open_mouth:oral:star:code:house:project:tv:video
• VR/AR
  • The Power of Points for Modeling Humans in Clothing
    :star:code:house:project:tv:video
  • 虚拟试穿
    • M3D-VTON: A Monocular-to-3D Virtual Try-On Network
      :star:code
    • ZFlow: Gated Appearance Flow-based Virtual Try-on with 3D Priors
    • Dressing in Order: Recurrent Person Image Generation for Pose Transfer, Virtual Try-On and Outfit Editing
    • FashionMirror: Co-Attention Feature-Remapping Virtual Try-On With Sequential Template Poses
    • Structure-transformed Texture-enhanced Network for Person Image Synthesis
• SLAM
  • On the Limits of Pseudo Ground Truth in Visual Camera Re-localisation
    :star:code
  • Transfusion: A Novel SLAM Method Focused on Transparent Objects
  • iMAP: Implicit Mapping and Positioning in Real-Time
  • Learning To Bundle-Adjust: A Graph Network Approach to Faster Optimization of Bundle Adjustment for Vehicular SLAM
  • R-SLAM: Optimizing Eye Tracking From Rolling Shutter Video of the Retina
  • Place Recognition
    • Attentional Pyramid Pooling of Salient Visual Residuals for Place Recognition
    • Pyramid Point Cloud Transformer for Large-Scale Place Recognition
      :star:code

35.Quantization/Pruning/Knowledge Distillation/Model Compression(量化、剪枝、蒸馏、模型压缩/扩展与优化)

• 知识蒸馏
  • Distilling Holistic Knowledge with Graph Neural Networks
    :star:code
  • Lipschitz Continuity Guided Knowledge Distillation
    :star:code
  • Densely Guided Knowledge Distillation Using Multiple Teacher Assistants
    :star:code
  • Revisiting Adversarial Robustness Distillation: Robust Soft Labels Make Student Better
    :star:code
  • Compressing Visual-linguistic Model via Knowledge Distillation
  • Self-Knowledge Distillation With Progressive Refinement of Targets
    :star:code:tv:video
  • Student Customized Knowledge Distillation: Bridging the Gap Between Student and Teacher
  • Channel-Wise Knowledge Distillation for Dense Prediction
    :star:code
  • Exploring Inter-Channel Correlation for Diversity-Preserved Knowledge Distillation
    :star:code
• 量化
  • Distance-aware Quantization
    :star:code:house:project
  • Dynamic Network Quantization for Efficient Video Inference
    :star:code:house:project
  • Cluster-Promoting Quantization with Bit-Drop for Minimizing Network Quantization Loss
  • Improving Low-Precision Network Quantization via Bin Regularization
  • Towards Mixed-Precision Quantization of Neural Networks via Constrained Optimization
  • Integer-arithmetic-only Certified Robustness for Quantized Neural Networks
  • RMSMP: A Novel Deep Neural Network Quantization Framework With Row-Wise Mixed Schemes and Multiple Precisions
  • Improving Neural Network Efficiency via Post-Training Quantization With Adaptive Floating-Point
    :star:code
  • Once Quantization-Aware Training: High Performance Extremely Low-bit Architecture Search
    :star:code
• 模型压缩
  • GDP: Stabilized Neural Network Pruning via Gates with Differentiable Polarization
  • Exploration and Estimation for Model Compression
• 剪枝
  • ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgetting
    :star:code
  • Auto Graph Encoder-Decoder for Neural Network Pruning

34.Super-Resolution(超分辨率)

• ISR
  • Mutual Affine Network for Spatially Variant Kernel Estimation in Blind Image Super-Resolution
    :star:code
  • Hierarchical Conditional Flow: A Unified Framework for Image Super-Resolution and Image Rescaling
    :star:code
  • Deep Reparametrization of Multi-Frame Super-Resolution and Denoising
    :open_mouth:oral
  • Dual-Camera Super-Resolution with Aligned Attention Modules
    :star:code:house:project:tv:video
  • Attention-Based Multi-Reference Learning for Image Super-Resolution
    :star:code:house:project
  • Learning a Single Network for Scale-Arbitrary Super-Resolution
  • Fourier Space Losses for Efficient Perceptual Image Super-Resolution
    :star:code
  • Achieving On-Mobile Real-Time Super-Resolution With Neural Architecture and Pruning Search
  • Designing a Practical Degradation Model for Deep Blind Image Super-Resolution
    :star:code
  • Event Stream Super-Resolution via Spatiotemporal Constraint Learning
  • Dynamic High-Pass Filtering and Multi-Spectral Attention for Image Super-Resolution
  • Super-Resolving Cross-Domain Face Miniatures by Peeking at One-Shot Exemplar
  • Context Reasoning Attention Network for Image Super-Resolution
  • EvIntSR-Net: Event Guided Multiple Latent Frames Reconstruction and Super-Resolution
  • Super Resolve Dynamic Scene from Continuous Spike Streams
  • Deep Blind Video Super-Resolution
  • Benchmarking Ultra-High-Definition Image Super-Resolution
  • Lucas-Kanade Reloaded: End-to-End Super-Resolution From Raw Image Bursts
  • Unsupervised Real-World Super-Resolution: A Domain Adaptation Perspective
  • Real-World Video Super-Resolution: A Benchmark Dataset and a Decomposition Based Learning Scheme
    :star:code
    :newspaper:解读:ICCV2021 香港理工、阿里达摩院提出RealVSR：视频超分任务中的新数据集与损失方案
• VSR
  • Omniscient Video Super-Resolution
    :star:code
  • COMISR: Compression-Informed Video Super-Resolution
    :star:code
    :newspaper:解读:谷歌提出COMISR算法：针对视频压缩的压缩感知超分辨率
  • Learning Frequency-Aware Dynamic Network for Efficient Super-Resolution
  • Efficient Video Compression via Content-Adaptive Super-Resolution
    :star:code

33.Remote Sensing Images(遥感影像)

• SUNet: Symmetric Undistortion Network for Rolling Shutter Correction
• Change is Everywhere: Single-Temporal Supervised Object Change Detection in Remote Sensing Imagery
  :star:code
  :newspaper:解读:ICCV2021｜武汉大学RSIDEA团队提出一种新颖的弱监督遥感变化检测算法STAR
• 卫星图像全景视频合成
  • Sat2Vid: Street-view Panoramic Video Synthesis from a Single Satellite Image
• 基于卫星影像的交通事故检测
  • Inferring High-Resolution Traffic Accident Risk Maps Based on Satellite Imagery and GPS Trajectories
• 遥感数据
  • Seasonal Contrast: Unsupervised Pre-Training From Uncurated Remote Sensing Data
    :star:code
  • Dynamic Cross Feature Fusion for Remote Sensing Pansharpening
• 分割
  • Self-Mutating Network for Domain Adaptive Segmentation in Aerial Images
  • 卫星图像的全景分割
    • Panoptic Segmentation of Satellite Image Time Series With Convolutional Temporal Attention Networks
      :star:code:sunflower:PASTIS dataset
• 三维重建
  • 3D Building Reconstruction from Monocular Remote Sensing Images
    :house:project

32.语音

• The Right to Talk: An Audio-Visual Transformer Approach
  :star:code
• Image2Reverb: Cross-Modal Reverb Impulse Response Synthesis
  :star:code:house:project
• 音频分离
  • Visual Scene Graphs for Audio Source Separation
• 音频-手势
  • Audio2Gestures: Generating Diverse Gestures from Speech Audio with Conditional Variational Autoencoders
    :house:project
• Active Speaker Detection(ASD主动式扬声器检测)
  • How To Design a Three-Stage Architecture for Audio-Visual Active Speaker Detection in the Wild
    :star:code
  • MAAS: Multi-Modal Assignation for Active Speaker Detection
• 从人脸视频中重新收集音频
  • Multi-Modality Associative Bridging Through Memory: Speech Sound Recollected From Face Video
• 视听源定位
  • Localize to Binauralize: Audio Spatialization From Visual Sound Source Localization
    :star:code:tv:video
• 视听源分离
  • Move2Hear: Active Audio-Visual Source Separation
    :star:code:house:project
• 视听平面图重建
  • Audio-Visual Floorplan Reconstruction
    :star:code:house:project:tv:video

31.Style Transfer(风格迁移)

• AdaAttN: Revisit Attention Mechanism in Arbitrary Neural Style Transfer
  :star:code
• Domain-Aware Universal Style Transfer
  :star:code
• Diverse Image Style Transfer via Invertible Cross-Space Mapping
• StyleFormer: Real-Time Arbitrary Style Transfer via Parametric Style Composition
• Manifold Alignment for Semantically Aligned Style Transfer
  :star:code

30.Image Generation/synthesis(图像生成/合成)

• ILVR: Conditioning Method for Denoising Diffusion Probabilistic Models
  :open_mouth:oral
• Image Synthesis via Semantic Composition
  :star:code:house:project
• Image Synthesis From Layout With Locality-Aware Mask Adaption
• 图像融合
  • DTMNet: A Discrete Tchebichef Moments-Based Deep Neural Network for Multi-Focus Image Fusion

29.Image Retrieval(图像检索)

• DOLG: Single-Stage Image Retrieval with Deep Orthogonal Fusion of Local and Global Features
  :star:code
• Image Retrieval on Real-life Images with Pre-trained Vision-and-Language Models
  :star:code:house:project
• Self-supervised Product Quantization for Deep Unsupervised Image Retrieval
  :star:code
• Instance-Level Image Retrieval Using Reranking Transformers
  :star:code
• Learning Attribute-Driven Disentangled Representations for Interactive Fashion Retrieval
  :star:code
• Telling the What While Pointing to the Where: Multimodal Queries for Image Retrieval
• Frozen in Time: A Joint Video and Image Encoder for End-to-End Retrieval
• Learning Deep Local Features With Multiple Dynamic Attentions for Large-Scale Image Retrieval
  :star:code
• Bayesian Triplet Loss: Uncertainty Quantification in Image Retrieval
• 跨域检索
  • Universal Cross-Domain Retrieval: Generalizing Across Classes and Domains
• Visual Geolocalization
  • Viewpoint Invariant Dense Matching for Visual Geolocalization
    :star:code
• 跨模态检索
  • Ask&Confirm: Active Detail Enriching for Cross-Modal Retrieval With Partial Query
    :star:code
  • Product1M: Towards Weakly Supervised Instance-Level Product Retrieval via Cross-modal Pretraining
    :star:code
  • Wasserstein Coupled Graph Learning for Cross-Modal Retrieval
  • Adversarial Attack on Deep Cross-Modal Hamming Retrieval
• 文本-视频检索
  • TEACHTEXT: CrossModal Generalized Distillation for Text-Video Retrieval
    :house:project
• 视频- 文本检索
  • HiT: Hierarchical Transformer With Momentum Contrast for Video-Text Retrieval
• image-based 3D shape retrieval
  • Single Image 3D Shape Retrieval via Cross-Modal Instance and Category Contrastive Learning
• 近邻搜索
  • Product Quantizer Aware Inverted Index for Scalable Nearest Neighbor Search

28.Contrastive Learning(对比学习)

• Improving Contrastive Learning by Visualizing Feature Transformation
  :open_mouth:oral:star:code
• TACo: Token-aware Cascade Contrastive Learning for Video-Text Alignment
  :newspaper:解读:ICCV2021-TOCo-微软&CMU提出Token感知的级联对比学习方法，在视频文本对齐任务上“吊打”其他SOTA方法
• A Broad Study on the Transferability of Visual Representations With Contrastive Learning
  :star:code
• Vi2CLR: Video and Image for Visual Contrastive Learning of Representation
• LatentCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable Directions
  :star:code
• CrossCLR: Cross-Modal Contrastive Learning for Multi-Modal Video Representations
• Social NCE: Contrastive Learning of Socially-Aware Motion Representations
  :star:code:tv:video
• With a Little Help From My Friends: Nearest-Neighbor Contrastive Learning of Visual Representations
• Contrastive Learning of Image Representations With Cross-Video Cycle-Consistency
  :house:project
• Weakly Supervised Contrastive Learning

27.Multi-label image recognition(多标签图像识别)

• Residual Attention: A Simple but Effective Method for Multi-Label Recognition
  :star:code
• Transformer-based Dual Relation Graph for Multi-label Image Recognition

26.Image Processing(图像处理)

• Aligning Latent and Image Spaces to Connect the Unconnectable
  :star:code:house:project
• 图像形状操纵
  • Image Shape Manipulation from a Single Augmented Training Sample
    :open_mouth:oral:star:code:house:project
• 边缘检测
  • RINDNet: Edge Detection for Discontinuity in Reflectance, Illumination, Normal and Depth
    :open_mouth:oral:star:code
  • Pixel Difference Networks for Efficient Edge Detection
    :star:code
• 图像识别
  • MicroNet: Improving Image Recognition with Extremely Low FLOPs
    :star:code
• 图像去模糊
  • Rethinking Coarse-to-Fine Approach in Single Image Deblurring
    :star:code
  • Single Image Defocus Deblurring Using Kernel-Sharing Parallel Atrous Convolutions
  • Defocus Map Estimation and Deblurring From a Single Dual-Pixel Image
  • Motion Deblurring with Real Events
  • Pyramid Architecture Search for Real-Time Image Deblurring
  • 运动去模糊
    • Perceptual Variousness Motion Deblurring With Light Global Context Refinement
• 视频去模糊
  • Bringing Events Into Video Deblurring With Non-Consecutively Blurry Frames
    :star:code
• Image quality assessment(图像质量评估IQA)
  • MUSIQ: Multi-scale Image Quality Transformer
    :star:code
• Image Harmonization
  • SSH: A Self-Supervised Framework for Image Harmonization
    :star:code
  • Learning Conditional Knowledge Distillation for Degraded-Reference Image Quality Assessment
    :star:code
• 去阴影
  • CANet: A Context-Aware Network for Shadow Removal
    :star:code
  • DC-ShadowNet: Single-Image Hard and Soft Shadow Removal Using Unsupervised Domain-Classifier Guided Network
• 去噪
  • Rethinking Deep Image Prior for Denoising
    :star:code
  • Rethinking Noise Synthesis and Modeling in Raw Denoising
    :star:code
  • C2N: Practical Generative Noise Modeling for Real-World Denoising
  • The Benefit of Distraction: Denoising Camera-Based Physiological Measurements Using Inverse Attention
    :star:code
  • Hyperspectral Image Denoising with Realistic Data
    :star:code
  • End-to-End Unsupervised Document Image Blind Denoising
  • Cross-Patch Graph Convolutional Network for Image Denoising
  • 视频去噪
    • Patch Craft: Video Denoising by Deep Modeling and Patch Matching
• 图像着色
  • Towards Vivid and Diverse Image Colorization with Generative Color Prior
  • Deep Edge-Aware Interactive Colorization Against Color-Bleeding Effects
    :open_mouth:oral:house:project
• 图像增强
  • Real-time Image Enhancer via Learnable Spatial-aware 3D Lookup Tables
  • Adaptive Unfolding Total Variation Network for Low-Light Image Enhancement
    :star:code
  • Representative Color Transform for Image Enhancement
  • STAR: A Structure-Aware Lightweight Transformer for Real-Time Image Enhancement
  • Deep Symmetric Network for Underexposed Image Enhancement With Recurrent Attentional Learning
    :star:code:house:project
  • StarEnhancer: Learning Real-Time and Style-Aware Image Enhancement
• 图像恢复
  • Spatially-Adaptive Image Restoration using Distortion-Guided Networks
    :star:code
  • Dynamic Attentive Graph Learning for Image Restoration
    :star:code
  • Self-Supervised Cryo-Electron Tomography Volumetric Image Restoration From Single Noisy Volume With Sparsity Constraint
    :star:code
  • Searching for Controllable Image Restoration Networks
    :star:code
• 图像压缩
  • Variable-Rate Deep Image Compression through Spatially-Adaptive Feature Transform
    :star:code
  • Neural Image Compression via Attentional Multi-Scale Back Projection and Frequency Decomposition
• 图像修复
  • Image Inpainting via Conditional Texture and Structure Dual Generation
    :star:code
  • CR-Fill: Generative Image Inpainting With Auxiliary Contextual Reconstruction
    :star:code
  • Parallel Multi-Resolution Fusion Network for Image Inpainting
  • Painting from Part
    :star:code
  • WaveFill: A Wavelet-Based Generation Network for Image Inpainting
  • Distillation-Guided Image Inpainting
  • Learning a Sketch Tensor Space for Image Inpainting of Man-made Scenes
    :star:code:house:project
• Image extrapolation
  • SemIE: Semantically-aware Image Extrapolation
    :house:project
• Reversible Image Conversion
  • IICNet: A Generic Framework for Reversible Image Conversion
    :star:code
• 伪影去除
  • Towards Flexible Blind JPEG Artifacts Removal
    :star:code
  • Learning Dual Priors for JPEG Compression Artifacts Removal
  • Let's See Clearly: Contaminant Artifact Removal for Moving Cameras
• De-rendering
  • De-rendering Stylized Texts
    :star:code:house:project
• 去除光晕
  • Light Source Guided Single-Image Flare Removal From Unpaired Data
• 全景图拼接
  • Minimal Solutions for Panoramic Stitching Given Gravity Prior
• Flare Removal
  • How to Train Neural Networks for Flare Removal
    :house:project:tv:video
• 图像裁剪
  • TransView: Inside, Outside, and Across the Cropping View Boundaries
  • Dissecting Image Crops
    :star:code
• 去反射
  • Location-Aware Single Image Reflection Removal
    :star:code
  • V-DESIRR: Very Fast Deep Embedded Single Image Reflection Removal
    :star:code
• 去雨
  • Improving De-Raining Generalization via Neural Reorganization
  • Unpaired Learning for Deep Image Deraining with Rain Direction Regularize
    :house:project
  • Structure-Preserving Deraining With Residue Channel Prior Guidance
    :star:code
• 图像失真去除
  • Unsupervised Non-Rigid Image Distortion Removal via Grid Deformation
    :star:code:tv:video
• 消除水下图像的折射失真
  • Learning To Remove Refractive Distortions From Underwater Images
• 图像补全
  • High-Fidelity Pluralistic Image Completion With Transformers
    :star:code:house:project
• Image Decomposition
  • Unsupervised Layered Image Decomposition into Object Prototypes
• 失真矫正
  • Towards Complete Scene and Regular Shape for Distortion Rectification by Curve-Aware Extrapolation
• HDR
  • Unpaired Learning for High Dynamic Range Image Tone Mapping
  • 超高清图像HDR重建
    • Ultra-High-Definition Image HDR Reconstruction via Collaborative Bilateral Learning
• 图像去雪
  • ALL Snow Removed: Single Image Desnowing Algorithm Using Hierarchical Dual-Tree Complex Wavelet Representation and Contradict Channel Loss
    :star:code
• Image Harmonization
  • Image Harmonization With Transformer
    :star:code
• 图像编辑
  • Language-Guided Global Image Editing via Cross-Modal Cyclic Mechanism
• image hiding(图像隐藏)
  • HiNet: Deep Image Hiding by Invertible Network
    :star:code

25.Medical Image(医学影像)

• Equivariant Imaging: Learning Beyond the Range Space
  :open_mouth:oral:star:code
• Deep Survival Analysis With Longitudinal X-Rays for COVID-19
• 医学图像分割
  • Recurrent Mask Refinement for Few-Shot Medical Image Segmentation
    :star:code
  • Graph-BAS3Net: Boundary-Aware Semi-Supervised Segmentation Network With Bilateral Graph Convolution医学图像分割
  • 病变分割
    • T-AutoML: Automated Machine Learning for Lesion Segmentation Using Transformers in 3D Medical Imaging
  • 息肉分割
    • Collaborative and Adversarial Learning of Focused and Dispersive Representations for Semi-Supervised Polyp Segmentation
  • 血管分割
    • Self-Supervised Vessel Segmentation via Adversarial Learning
  • 脑肿瘤分割
    • RFNet: Region-Aware Fusion Network for Incomplete Multi-Modal Brain Tumor Segmentation
• 病理学图像表示
  • A QuadTree Image Representation for Computational Pathology
• 医学图像分析
  • Preservational Learning Improves Self-supervised Medical Image Models by Reconstructing Diverse Contexts
    :star:code
    :newspaper:解读:ICCV2021 2D和3D通用！新医疗影像自监督SOTA（代码已开源）
• 医学图像去噪
  • Eformer: Edge Enhancement based Transformer for Medical Image Denoising
• 视频翻译
  • Long-Term Temporally Consistent Unpaired Video Translation From Simulated Surgical 3D Data
    :star:code:house:project
• 病理学图像核检测分割
  • Mutual-Complementing Framework for Nuclei Detection and Segmentation in Pathology Image
• 医学报告生成
  • Visual-Textual Attentive Semantic Consistency for Medical Report Generation
• CT
  • 3DeepCT: Learning Volumetric Scattering Tomography of Clouds
  • IntraTomo: Self-Supervised Learning-Based Tomography via Sinogram Synthesis and Prediction
    :star:code
  • CT重建
    • Dynamic CT Reconstruction From Limited Views With Implicit Neural Representations and Parametric Motion Fields
      :star:code
• 医学图像识别
  • GLoRIA: A Multimodal Global-Local Representation Learning Framework for Label-Efficient Medical Image Recognition
    :star:code
• 医学图像分类
  • Big Self-Supervised Models Advance Medical Image Classification
  • Large-Scale Robust Deep AUC Maximization: A New Surrogate Loss and Empirical Studies on Medical Image Classification
    :star:code

24.Face(人脸)

• VariTex: Variational Neural Face Textures
  :star:code:house:project:tv:video
• 人脸造假检测
  • OpenForensics: Large-Scale Challenging Dataset For Multi-Face Forgery Detection And Segmentation In-The-Wild
    :house:project
  • Exploring Temporal Coherence for More General Video Face Forgery Detection
• 人脸合成
  • Disentangled Lifespan Face Synthesis
    :star:code:house:project:tv:video
• 人脸识别
  • PASS: Protected Attribute Suppression System for Mitigating Bias in Face Recognition
  • SynFace: Face Recognition with Synthetic Data
  • Adaptive Label Noise Cleaning With Meta-Supervision for Deep Face Recognition
  • Disentangled Representation for Age-Invariant Face Recognition: A Mutual Information Minimization Perspective
  • Teacher-Student Adversarial Depth Hallucination To Improve Face Recognition
    :star:code
  • DAM: Discrepancy Alignment Metric for Face Recognition
  • “去”识别
    • Personalized and Invertible Face De-Identification by Disentangled Identity Information Manipulation
• Face perception面部感知
  • Learning Facial Representations from the Cycle-consistency of Face
    :star:code
• 说话人脸生成
  • FACIAL: Synthesizing Dynamic Talking Face with Implicit Attribute Learning
• 说话头合成
  • AD-NeRF: Audio Driven Neural Radiance Fields for Talking Head Synthesis
    :star:code
  • Learned Spatial Representations for Few-Shot Talking-Head Synthesis
    :star:code:house:project
• 人脸表情识别
  • Understanding and Mitigating Annotation Bias in Facial Expression Recognition
  • TransFER: Learning Relation-aware Facial Expression Representations with Transformers
• 人脸呈现攻击检测
  • Detection and Continual Learning of Novel Face Presentation Attacks
    :star:code
• 人脸编辑
  • Talk-to-Edit: Fine-Grained Facial Editing via Dialog
    :star:code:house:project
    :newspaper:解读:ICCV2021 | 南洋理工大学、港中大提出Talk-to-Edit，对话实现高细粒度人脸编辑
  • A Latent Transformer for Disentangled Face Editing in Images and Videos
    :star:code
• 人脸对齐
  • ADNet: Leveraging Error-Bias Towards Normal Direction in Face Alignment
• 人脸图像重建
  • Focal Frequency Loss for Image Reconstruction and Synthesis
    :star:code:house:project:tv:video
  • Towards High Fidelity Monocular Face Reconstruction With Rich Reflectance Using Self-Supervised Learning and Ray Tracing
  • Neural Photofit: Gaze-Based Mental Image Reconstruction
    :house:project
• 3D人脸重建
  • Topologically Consistent Multi-View Face Inference Using Volumetric Sampling
    :star:code
  • Self-Supervised 3D Face Reconstruction via Conditional Estimation
• 三维人脸动画
  • MeshTalk: 3D Face Animation From Speech Using Cross-Modality Disentanglement
    :star:code:tv:video
• Remote Photoplethysmography (rPPG远程光电容积描记术)
  • The Way to My Heart Is Through Contrastive Learning: Remote Photoplethysmography From Unlabelled Video
    :star:code
• 人脸加密
  • Towards Face Encryption by Generating Adversarial Identity Masks
    :star:code
• Deepfake检测
  • Learning Self-Consistency for Deepfake Detection
    :open_mouth:oral
  • Joint Audio-Visual Deepfake Detection
  • Artificial Fingerprinting for Generative Models: Rooting Deepfake Attribution in Training Data
    :open_mouth:oral
• 人脸纹理补全
  • Learning High-Fidelity Face Texture Completion Without Complete Face Texture
• 面部动作单元检测
  • PIAP-DF: Pixel-Interested and Anti Person-Specific Facial Action Unit Detection Net With Discrete Feedback Learning
• 人脸分析
  • Fake It Till You Make It: Face analysis in the wild using synthetic data alone
    :house:project:tv:video
• 3D头重建
  • H3D-Net: Few-Shot High-Fidelity 3D Head Reconstruction
• 人脸关键点检测
  • Improving Robustness of Facial Landmark Detection by Defending Against Adversarial Attacks
    :star:code
• 人脸图像检索
  • Face Image Retrieval with Attribute Manipulation

23.Gaze Estimation(视线估计)

• Generalizing Gaze Estimation with Outlier-guided Collaborative Adaptation
  :star:code
• 视线跟踪
  • Looking Here or There? Gaze Following in 360-Degree Images
• 视点估计
  • ViewNet: Unsupervised Viewpoint Estimation From Conditional Generation

22.GAN

• Sketch Your Own GAN
  :star:code:house:project
• Online Multi-Granularity Distillation for GAN Compression
  :star:code
• Dual Projection Generative Adversarial Networks for Conditional Image Generation
  :star:code
• InSeGAN: A Generative Approach to Segmenting Identical Instances in Depth Images
• ReStyle: A Residual-Based StyleGAN Encoder via Iterative Refinement
  :star:code:house:project:tv:video
• WarpedGANSpace: Finding non-linear RBF paths in GAN latent space
  :star:code
• Toward a Visual Concept Vocabulary for GAN Latent Space
• Collaging Class-specific GANs for Semantic Image Synthesis
  :house:project
• Latent Transformations via NeuralODEs for GAN-Based Image Editing
• Reality Transform Adversarial Generators for Image Splicing Forgery Detection and Localization
• GAN-Control: Explicitly Controllable GANs(https://alonshoshan10.github.io/gan_control/)
  :house:project
• Omni-GAN: On the Secrets of cGANs and Beyond
  :star:code
• Unsupervised Image Generation with Infinite Generative Adversarial Networks
  :star:code
• DAE-GAN: Dynamic Aspect-Aware GAN for Text-to-Image Synthesis
• Detail Me More: Improving GAN’s photo-realism of complex scenes
• Unsupervised Segmentation Incorporating Shape Prior via Generative Adversarial Networks
• DRB-GAN: A Dynamic ResBlock Generative Adversarial Network for Artistic Style Transfer
  :star:code
• Dual Contrastive Loss and Attention for GANs
• Semi-Supervised Single-Stage Controllable GANs for Conditional Fine-Grained Image Generation
• Gradient Normalization for Generative Adversarial Networks
  :star:code
• EigenGAN: Layer-Wise Eigen-Learning for GANs
  :star:code
• Retrieve in Style: Unsupervised Facial Feature Transfer and Retrieval
  :star:code
• HeadGAN: One-shot Neural Head Synthesis and Editing
  :house:project:tv:video
• Explaining in Style: Training a GAN To Explain a Classifier in StyleSpace
  :star:code:house:project:tv:video
• StyleCLIP: Text-Driven Manipulation of StyleGAN Imagery
  :open_mouth:oral:star:code:tv:video
• Towards Discovery and Attribution of Open-World GAN Generated Images
• Diagonal Attention and Style-Based GAN for Content-Style Disentanglement in Image Generation and Translation
• Re-Aging GAN: Toward Personalized Face Age Transformation
• When do GANs replicate? On the choice of dataset size
  :star:code
• LoFGAN: Fusing Local Representations for Few-shot Image Generation
• Multi-Class Multi-Instance Count Conditioned Adversarial Image Generation
  :star:code
• Generative Adversarial Registration for Improved Conditional Deformable Templates
  :star:code
• F-Drop&Match: GANs with a Dead Zone in the High-Frequency Domain
• GAN inversion(GAN逆映射)
  • From Continuity to Editability: Inverting GANs with Consecutive Images
    :star:code
  • GAN Inversion for Out-of-Range Images with Geometric Transformations
    :house:project
• 图像到图像翻译
  • Unaligned Image-to-Image Translation by Learning to Reweight
    :star:code
  • Bridging the Gap between Label- and Reference-based Synthesis in Multi-attribute Image-to-Image Translation
    :star:code
  • Instance-Wise Hard Negative Example Generation for Contrastive Learning in Unpaired Image-to-Image Translation
  • TransferI2I: Transfer Learning for Image-to-Image Translation from Small Datasets
    :star:code
  • Rethinking the Truly Unsupervised Image-to-Image Translation
  • SPatchGAN: A Statistical Feature Based Discriminator for Unsupervised Image-to-Image Translation
• Image translation
  • Scaling-up Disentanglement for Image Translation
    :star:code:house:project
  • Harnessing the Conditioning Sensorium for Improved Image Translation
  • Frequency Domain Image Translation: More Photo-Realistic, Better Identity-Preserving
    :star:code
  • Dual Transfer Learning for Event-based End-task Prediction via Pluggable Event to Image Translation
    :star:code
  • Semantically Robust Unpaired Image Translation for Data with Unmatched Semantics Statistics

21.Active Learning(主动学习)

• Semi-Supervised Active Learning with Temporal Output Discrepancy
  :star:code
• Influence Selection for Active Learning
• Active Domain Adaptation via Clustering Uncertainty-weighted Embeddings
  :star:code
• Contrastive Coding for Active Learning under Class Distribution Mismatch
  :star:code

20.Adversarial Learning(对抗学习)

• Low Curvature Activations Reduce Overfitting in Adversarial Training
• Removing Adversarial Noise in Class Activation Feature Space
• Sample Efficient Detection and Classification of Adversarial Attacks via Self-Supervised Embeddings
• Invisible Backdoor Attack With Sample-Specific Triggers
  :star:code
• Defending Against Universal Adversarial Patches by Clipping Feature Norms
• 对抗攻击
  • Feature Importance-aware Transferable Adversarial Attacks
    :star:code
  • TkML-AP: Adversarial Attacks to Top-k Multi-Label Learning
    :star:code
  • Meta Gradient Adversarial Attack
  • AGKD-BML: Defense Against Adversarial Attack by Attention Guided Knowledge Distillation and Bi-directional Metric Learning
    :star:code
  • Exploiting Multi-Object Relationships for Detecting Adversarial Attacks in Complex Scenes
  • AdvDrop: Adversarial Attack to DNNs by Dropping Information
    :star:code
  • Adversarial Attacks Are Reversible With Natural Supervision
  • Attack As the Best Defense: Nullifying Image-to-Image Translation GANs via Limit-Aware Adversarial Attack
  • Learnable Boundary Guided Adversarial Training
    :star:code
  • Augmented Lagrangian Adversarial Attacks
    :star:code
  • Meta-Attack: Class-Agnostic and Model-Agnostic Physical Adversarial Attack
  • On Generating Transferable Targeted Perturbations
    :star:code
  • Admix: Enhancing the Transferability of Adversarial Attacks
    :star:code
  • Consistency-Sensitivity Guided Ensemble Black-Box Adversarial Attacks in Low-Dimensional Spaces
  • Adversarial Attacks On Multi-Agent Communication
  • Interpreting Attributions and Interactions of Adversarial Attacks
  • RDA: Robust Domain Adaptation via Fourier Adversarial Attacking
• 对抗样本
  • Adversarial Example Detection Using Latent Neighborhood Graph
  • On the Robustness of Vision Transformers to Adversarial Examples
• 黑盒
  • Black-Box Detection of Backdoor Attacks With Limited Information and Data
  • Aha! Adaptive History-Driven Attack for Decision-Based Black-Box Models
  • Data-Free Universal Adversarial Perturbation and Black-Box Attack

19.Self-Driving Vehicles(自动驾驶)

• End-to-End Urban Driving by Imitating a Reinforcement Learning Coach
  :star:code
• MultiSiam: Self-supervised Multi-instance Siamese Representation Learning for Autonomous Driving
  :star:code
• NEAT: Neural Attention Fields for End-to-End Autonomous Driving
  :star:code
• Safety-aware Motion Prediction with Unseen Vehicles for Autonomous Driving
  :star:code
• Social-NCE: Contrastive Learning of Socially-aware Motion Representations
  :star:code:tv:video
• Learning To Drive From a World on Rails
  :open_mouth:oral:star:code:house:project
• DRIVE: Deep Reinforced Accident Anticipation With Visual Explanation
  :star:code:house:project:tv:video
• LookOut: Diverse Multi-Future Prediction and Planning for Self-Driving
• Prediction by Anticipation: An Action-Conditional Prediction Method Based on Interaction Learning
  :star:code:tv:video
• TMCOSS: Thresholded Multi-Criteria Online Subset Selection for Data-Efficient Autonomous Driving
• FIERY: Future Instance Prediction in Bird's-Eye View From Surround Monocular Cameras
  :star:code
• On Exposing the Challenging Long Tail in Future Prediction of Traffic Actors
  :star:code
• MGNet: Monocular Geometric Scene Understanding for Autonomous Driving
  :star:code:tv:video
• Human trajectory prediction(人体轨迹预测)
  • Personalized Trajectory Prediction via Distribution Discrimination
    :star:code
  • Human Trajectory Prediction via Counterfactual Analysis
    :star:code
  • From Goals, Waypoints & Paths to Long Term Human Trajectory Forecasting
    :star:code:house:project:tv:video
• 轨迹预测
  • Unlimited Neighborhood Interaction for Heterogeneous Trajectory Prediction
  • LOKI: Long Term and Key Intentions for Trajectory Prediction
  • MG-GAN: A Multi-Generator Model Preventing Out-of-Distribution Samples in Pedestrian Trajectory Prediction
    :star:code
  • DenseTNT: End-to-end Trajectory Prediction from Dense Goal Sets
  • Where Are You Heading? Dynamic Trajectory Prediction With Expert Goal Examples
    :star:code
  • Three Steps to Multimodal Trajectory Prediction: Modality Clustering, Classification and Synthesis
  • Spatial-Temporal Consistency Network for Low-Latency Trajectory Forecasting
  • Likelihood-Based Diverse Sampling for Trajectory Forecasting
    :star:code
• 运动预测
  • RAIN: Reinforced Hybrid Attention Inference Network for Motion Forecasting
    :house:project
  • SLAMP: Stochastic Latent Appearance and Motion Prediction
    :house:project
  • SlowFast Rolling-Unrolling LSTMs for Action Anticipation in Egocentric Videos
• 自动导航
  • FOVEA: Foveated Image Magnification for Autonomous Navigation
    :house:project
• 交通场景理解
  • Structured Bird's-Eye-View Traffic Scene Understanding from Onboard Images
    :star:code
• 车辆车牌识别
  • 车辆重识别
    • Heterogeneous Relational Complement for Vehicle Re-identification
    • Self-Supervised Geometric Features Discovery via Interpretable Attention for Vehicle Re-Identification and Beyond
• 自主赛车
  • Learn-to-Race: A Multimodal Control Environment for Autonomous Racing
    :star:code
• 预测司机的视觉注意力
  • MEDIRL: Predicting the Visual Attention of Drivers via Maximum Entropy Deep Inverse Reinforcement Learning
    :star:code
• 姿势预测
  • Space-Time-Separable Graph Convolutional Network for Pose Forecasting
    :star:code:tv:video
• 车辆跟踪
  • Track Without Appearance: Learn Box and Tracklet Embedding With Local and Global Motion Patterns for Vehicle Tracking
• 对任意相机视角的车辆进行检测分析
  • Robust 2D/3D Vehicle Parsing in Arbitrary Camera Views for CVIS
• 车道线检测
  • CondLaneNet: A Top-To-Down Lane Detection Framework Based on Conditional Convolution
    :star:code
  • Active Learning for Lane Detection: A Knowledge Distillation Approach
• 车速估计
  • Robust Automatic Monocular Vehicle Speed Estimation for Traffic Surveillance
    :star:code

18.Transformers

• Generic Attention-model Explainability for Interpreting Bi-Modal and Encoder-Decoder Transformers
  :open_mouth:oral:star:code
  :newspaper:解读:ICCV2021 Oral-TAU&Facebook提出了通用的Attention模型可解释性
• Transformer-Based Attention Networks for Continuous Pixel-Wise Prediction
  :star:code
• PlaneTR: Structure-Guided Transformers for 3D Plane Recovery
  :star:code
• Rethinking and Improving Relative Position Encoding for Vision Transformer
  :star:code
• Vision Transformer with Progressive Sampling
  :star:code
• Paint Transformer: Feed Forward Neural Painting with Stroke Prediction
  :open_mouth:oral:star:code
• Rethinking Spatial Dimensions of Vision Transformers
  :star:code
  :newspaper:解读:ICCV2021-PiT-池化操作不是CNN的专属，ViT说：“我也可以”；南大提出池化视觉Transformer（PiT）
• PnP-DETR: Towards Efficient Visual Analysis with Transformers
  :star:code
• Describing and Localizing Multiple Changes With Transformers
  :star:code:house:project
• LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference
  :star:code
• VidTr: Video Transformer Without Convolutions
• Visformer: The Vision-Friendly Transformer
  :star:code
• Going Deeper With Image Transformers
  :star:code
• Multiscale Vision Transformers
  :star:code
• Learning Multi-Scene Absolute Pose Regression With Transformers
  :star:code
• Visual Saliency Transformer
  :star:code
• Event-Based Video Reconstruction Using Transformer
  :star:code
• Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows
  :star:code
• An Empirical Study of Training Self-Supervised Vision Transformers
  :open_mouth:oral:star:code
• Tokens-to-Token ViT: Training Vision Transformers From Scratch on ImageNet
  :star:code
• CvT: Introducing Convolutions to Vision Transformers
  :star:code
• COTR: Correspondence Transformer for Matching Across Images
• ViViT: A Video Vision Transformer
  :star:code
• AgentFormer: Agent-Aware Transformers for Socio-Temporal Multi-Agent Forecasting
  :star:code:house:project
• Incorporating Convolution Designs into Visual Transformers
  :star:code
• LayoutTransformer: Layout Generation and Completion with Self-attention
  :star:code:house:project
• AutoFormer: Searching Transformers for Visual Recognition
  :star:code
• Scalable Vision Transformers With Hierarchical Pooling
  :star:code
• Visual Transformers: Where Do Transformers Really Belong in Vision Models?
• Anticipative Video Transformer
  :star:code:house:project
• 密集预测
  • Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions
    :open_mouth:oral:star:code
    :newspaper:解读:大白话Pyramid Vision Transformer
  • Vision Transformers for Dense Prediction
    :star:code
• 3D人体纹理估计
  • 3D Human Texture Estimation from a Single Image with Transformers
    :open_mouth:oral:house:project
• 图像编辑
  • Multi-Scale Vision Longformer: A New Vision Transformer for High-Resolution Image Encoding
    :star:code
• OCR
  • DocFormer: End-to-End Transformer for Document Understanding
• 根据音乐生成舞蹈
  • AI Choreographer: Music Conditioned 3D Dance Generation With AIST++
    :house:project
    :newspaper:简介:Transformer又又来了，生成配有音乐的丝滑3D舞蹈，开放最大规模数据集AIST++

17.3D(三维视觉)

• Discovering 3D Parts from Image Collections
  :open_mouth:oral:star:code:house:project:tv:video
• PixelSynth: Generating a 3D-Consistent Experience from a Single Image
  :star:code:house:project
• Towers of Babel: Combining Images, Language, and 3D Geometry for Learning Multimodal Vision
  :star:code:house:project
• Pixel-Perfect Structure-from-Motion with Featuremetric Refinement
  :star:code
• Learning Anchored Unsigned Distance Functions with Gradient Direction Alignment for Single-view Garment Reconstruction
  :open_mouth:oral:star:code
• LSD-StructureNet: Modeling Levels of Structural Detail in 3D Part Hierarchies
• Rational Polynomial Camera Model Warping for Deep Learning Based Satellite Multi-View Stereo Matching
  :star:code
• Where2Act: From Pixels to Actions for Articulated 3D Objects
  :tv:video
• BuildingNet: Learning to Label 3D Buildings
  :open_mouth:oral:star:code:house:project
• SurfGen: Adversarial 3D Shape Synthesis With Explicit Surface Discriminators
• Deep Virtual Markers for Articulated 3D Shapes
  :star:code:tv:video
• Learning Efficient Photometric Feature Transform for Multi-view Stereo
  :house:project
• Neural-GIF: Neural Generalized Implicit Functions for Animating People in Clothing
• Just a Few Points Are All You Need for Multi-View Stereo: A Novel Semi-Supervised Learning Method for Multi-View Stereo
• 3D-FRONT: 3D Furnished Rooms With layOuts and semaNTics
• Learning Generative Models of Textured 3D Meshes from Real-World Images
  :star:code
• Self-Supervised Pretraining of 3D Features on any Point-Cloud
• High Quality Disparity Remapping with Two-Stage Warping
• Structure-From-Sherds: Incremental 3D Reassembly of Axially Symmetric Pots From Unordered and Mixed Fragment Collections
  :star:code
• Interpolation-Aware Padding for 3D Sparse Convolutional Neural Networks
• 深度估计
  • StructDepth: Leveraging the structural regularities for self-supervised indoor depth estimation
    :star:code
  • Bridging Unsupervised and Supervised Depth from Focus via All-in-Focus Supervision
    :star:code:house:project
  • Augmenting Depth Estimation with Geospatial Context
  • Can Scale-Consistent Monocular Depth Be Learned in a Self-Supervised Scale-Invariant Manner?
  • Revisiting Stereo Depth Estimation From a Sequence-to-Sequence Perspective With Transformers
    :open_mouth:oral:star:code
  • Adaptive Surface Normal Constraint for Depth Estimation
    :star:code
  • Event-Intensity Stereo: Estimating Depth by the Best of Both Worlds
  • DnD: Dense Depth Estimation in Crowded Dynamic Indoor Scenes
  • DepthInSpace: Exploitation and Fusion of Multiple Video Frames for Structured-Light Depth Estimation
    :house:project
  • Boosting Monocular Depth Estimation With Lightweight 3D Point Fusion
  • Monocular Depth Estimation(单目深度估计)
    • Revealing the Reciprocal Relations Between Self-Supervised Stereo and Monocular Depth Estimation
    • MonoIndoor: Towards Good Practice of Self-Supervised Monocular Depth Estimation for Indoor Environments
    • Regularizing Nighttime Weirdness: Efficient Self-supervised Monocular Depth Estimation in the Dark
      :star:code
    • Towards Interpretable Deep Networks for Monocular Depth Estimation
      :star:code:tv:video
    • Self-supervised Monocular Depth Estimation for All Day Images using Domain Separation
      :star:code
    • Fine-grained Semantics-aware Representation Enhancement for Self-supervised Monocular Depth Estimation
      :open_mouth:oral:star:code
    • Excavating the Potential Capacity of Self-Supervised Monocular Depth Estimation
      :star:code
    • R-MSFM: Recurrent Multi-Scale Feature Modulation for Monocular Depth Estimating
      :star:code
    • Adaptive Confidence Thresholding for Monocular Depth Estimation(https://github.com/megvii-research/OMNet)
    • SaccadeCam: Adaptive Visual Attention for Monocular Depth Sensing
• 深度补全
  • Bayesian Deep Basis Fitting for Depth Completion With Uncertainty
  • Unsupervised Depth Completion With Calibrated Backprojection Layers
    :star:code
• Omnidirectional Localization
  • PICCOLO: Point Cloud-Centric Omnidirectional Localization
• 三维重建
  • Learning Signed Distance Field for Multi-view Surface Reconstruction
    :open_mouth:oral
  • 3DStyleNet: Creating 3D Shapes with Geometric and Texture Style Variations
    :open_mouth:oral:house:project
  • DensePose 3D: Lifting Canonical Surface Maps of Articulated Objects to the Third Dimension
  • In-the-Wild Single Camera 3D Reconstruction Through Moving Water Surfaces
    :open_mouth:oral:star:code:tv:video
  • Gaussian Fusion: Accurate 3D Reconstruction via Geometry-Guided Displacement Interpolation
  • RetrievalFuse: Neural 3D Scene Reconstruction With a Database
    :star:code:house:project:tv:video
  • Multi-View 3D Reconstruction With Transformers
  • Polarimetric Helmholtz Stereopsis
  • MINE: Towards Continuous Depth MPI with NeRF for Novel View Synthesis
    :star:code:tv:video
  • Toward Realistic Single-View 3D Object Reconstruction With Unsupervised Learning From Multiple Images
    :star:code
  • CryoDRGN2: Ab initio neural reconstruction of 3D protein structures from real cryo-EM images
  • 三维场景重建
    • Graph-to-3D: End-to-End Generation and Manipulation of 3D Scenes Using Scene Graphs
    • VolumeFusion: Deep Depth Fusion for 3D Scene Reconstruction
  • 三维形状重建
    • 3DIAS: 3D Shape Reconstruction with Implicit Algebraic Surfaces
      :house:project
    • Multiresolution Deep Implicit Functions for 3D Shape Representation
    • MVTN: Multi-View Transformation Network for 3D Shape Recognition
      :star:code:tv:video
    • Sketch2Mesh: Reconstructing and Editing 3D Shapes from Sketches
  • 三维网格重建
    • Vis2Mesh: Efficient Mesh Reconstruction from Unstructured Point Clouds of Large Scenes with Learned Virtual View Visibility
      :star:code
• 三维场景
  • Scene Synthesis via Uncertainty-Driven Attribute Synchronization
    :star:code
  • Graph-to-3D: End-to-End Generation and Manipulation of 3D Scenes Using Scene Graphs
    :house:project
• 相机校准
  • CTRL-C: Camera calibration TRansformer with Line-Classification
    :star:code
  • BabelCalib: A Universal Approach to Calibrating Central Cameras
    :star:code
• 表面重建
  • Temporally-Coherent Surface Reconstruction via Metric-Consistent Atlases
  • Planar Surface Reconstruction from Sparse Views
    :open_mouth:oral:star:code:house:project:tv:video
  • Adaptive Surface Reconstruction With Multiscale Convolutional Kernels
• 3D场景合成
  • Indoor Scene Generation From a Collection of Semantic-Segmented Depth Images
    :star:code
• 3D形状识别
  • Learning Canonical View Representation for 3D Shape Recognition With Arbitrary Views
    :star:code
• 图像重建
  • Semantic-embedded Unsupervised Spectral Reconstruction from Single RGB Images in the Wild
    :star:code
• Multi-view Stereo(MVS)
  • Digging into Uncertainty in Self-supervised Multi-view Stereo
    :star:code
  • PatchMatch-RL: Deep MVS with Pixelwise Depth, Normal, and Visibility
    :star:code
  • A Confidence-Based Iterative Solver of Depths and Surface Normals for Deep Multi-View Stereo
    :star:code
  • EPP-MVSNet: Epipolar-Assembling Based Depth Prediction for Multi-View Stereo

16.Re-Identification(重识别)

Object Re-Identification目标(物体)重识别

• TransReID: Transformer-Based Object Re-Identification
  :star:code

Person Re-Identification(人员重识别)

• Spatio-Temporal Representation Factorization for Video-based Person Re-Identification
• Learning Instance-level Spatial-Temporal Patterns for Person Re-identification
  :star:code
• Towards Discriminative Representation Learning for Unsupervised Person Re-identification
• Learning by Aligning: Visible-Infrared Person Re-identification using Cross-Modal Correspondences
  :star:code:house:project
• Video-based Person Re-identification with Spatial and Temporal Memory Networks
  :house:project
• Multi-Expert Adversarial Attack Detection in Person Re-identification Using Context Inconsistency
• Clothing Status Awareness for Long-Term Person Re-Identification
• Dense Interaction Learning for Video-Based Person Re-Identification
  :open_mouth:oral
• Explainable Person Re-Identification With Attribute-Guided Metric Distillation
  :house:project
• Online Pseudo Label Generation by Hierarchical Cluster Dynamics for Adaptive Person Re-Identification
• Pyramid Spatial-Temporal Aggregation for Video-Based Person Re-Identification
  :star:code
• ICE: Inter-Instance Contrastive Encoding for Unsupervised Person Re-Identification
  :star:code:tv:video
• Learning To Know Where To See: A Visibility-Aware Approach for Occluded Person Re-Identification
• Attack-Guided Perceptual Data Generation for Real-world Re-Identification
• BV-Person: A Large-Scale Dataset for Bird-View Person Re-Identification
  :sunflower:dataset
• CM-NAS: Cross-Modality Neural Architecture Search for Visible-Infrared Person Re-Identification
  :star:code
• Meta Pairwise Relationship Distillation for Unsupervised Person Re-Identification
• Syncretic Modality Collaborative Learning for Visible Infrared Person Re-Identification
• Weakly Supervised Text-Based Person Re-Identification
  :star:code
• Occlude Them All: Occlusion-Aware Attention Network for Occluded Person Re-ID
• Occluded Person Re-Identification with Single-scale Global Representations
  :star:code
• 域适应人员重识别
  • IDM: An Intermediate Domain Module for Domain Adaptive Person Re-ID
    :open_mouth:oral:star:code
• Crowd Counting(拥挤人群计数)
  • Rethinking Counting and Localization in Crowds:A Purely Point-Based Framework
    :open_mouth:oral:star:code
  • Uniformity in Heterogeneity:Diving Deep into Count Interval Partition for Crowd Counting
    :star:code
  • Spatial Uncertainty-Aware Semi-Supervised Crowd Counting
    :star:code
  • Variational Attention: Propagating Domain-Specific Knowledge for Multi-Domain Learning in Crowd Counting
    :star:code
  • Exploiting Sample Correlation for Crowd Counting With Multi-Expert Network
    :star:code
  • Crowd Counting With Partial Annotations in an Image
    :star:code
  • Towards A Universal Model for Cross-Dataset Crowd Counting
  • Rethinking Counting and Localization in Crowds:A Purely Point-Based Framework
    :open_mouth:oral:star:code
  • Uniformity in Heterogeneity: Diving Deep Into Count Interval Partition for Crowd Counting
    :star:code
• 跨模态人员重识别
  • Cross-Modality Person Re-Identification via Modality Confusion and Center Aggregation
• 行人检测
  • MOTSynth: How Can Synthetic Data Help Pedestrian Detection and Tracking?
  • Stacked Homography Transformations for Multi-View Pedestrian Detection
  • Robust Small-scale Pedestrian Detection with Cued Recall via Memory Learning
  • Body-Face Joint Detection via Embedding and Head Hook
    :star:code
• 行人属性识别
  • Spatial and Semantic Consistency Regularizations for Pedestrian Attribute Recognition
  • LapsCore: Language-Guided Person Search via Color Reasoning
• Person Search(行人搜索)
  • Weakly Supervised Person Search with Region Siamese Networks
  • End-to-End Trainable Trident Person Search Network Using Adaptive Gradient Propagation
  • ASMR: Learning Attribute-Based Person Search with Adaptive Semantic Margin Regularizer
    :house:project
  • Weakly Supervised Person Search with Region Siamese Networks
• 行人行为预测
  • Bifold and Semantic Reasoning for Pedestrian Behavior Prediction
• 步态识别
  • Gait Recognition via Effective Global-Local Feature Representation and Local Temporal Aggregation
  • Context-Sensitive Temporal Feature Learning for Gait Recognition
    :star:code
  • 3D Local Convolutional Neural Networks for Gait Recognition
    :star:code
  • Gait Recognition in the Wild: A Benchmark
    :star:code:house:project

15.Object Tracking(目标跟踪)

• Saliency-Associated Object Tracking
• Box-Aware Feature Enhancement for Single Object Tracking on Point Clouds
  :star:code
• Learning to Track Objects from Unlabeled Videos
  :star:code
• DepthTrack : Unveiling the Power of RGBD Tracking
  :star:code
• Learning Target Candidate Association To Keep Track of What Not To Track
  :star:code
• Transparent Object Tracking Benchmark
  :house:project
• DepthTrack: Unveiling the Power of RGBD Tracking
  :star:code
• Object Tracking by Jointly Exploiting Frame and Event Domain
• High-Performance Discriminative Tracking with Transformers
• Visio-Temporal Attention for Multi-Camera Multi-Target Association
• 视觉目标跟踪
  • Learning to Adversarially Blur Visual Object Tracking
    :star:code
  • Learn to Match: Automatic Matching Network Design for Visual Tracking
    :star:code
  • Video Annotation for Visual Tracking via Selection and Refinement
    :star:code
  • Learning Spatio-Temporal Transformer for Visual Tracking
    :star:code
  • 3D视觉跟踪
    • MLVSNet: Multi-Level Voting Siamese Network for 3D Visual Tracking
      :star:code
• 卫星图像跟踪
  • HiFT: Hierarchical Feature Transformer for Aerial Tracking
    :star:code
• 3D多目标跟踪
  • Exploring Simple 3D Multi-Object Tracking for Autonomous Driving
    :star:code
    :newspaper:解读:ICCV 2021丨轻舟智航提出SimTrack: 3D多目标一体化检测与跟踪，简单又精确
• 多目标跟踪与分割
  • Assignment-Space-Based Multi-Object Tracking and Segmentation
  • Continuous Copy-Paste for One-Stage Multi-Object Tracking and Segmentation
    :star:code
• 多目标跟踪
  • A General Recurrent Tracking Framework Without Real Data
  • Making Higher Order MOT Scalable: An Efficient Approximate Solver for Lifted Disjoint Paths
    :star:code
  • MOTSynth: How Can Synthetic Data Help Pedestrian Detection and Tracking?
  • Learning To Track With Object Permanence
    :star:code多目标跟踪
• 视频目标跟踪
  • TF-Blender: Temporal Feature Blender for Video Object Detection
    :star:code

14.Object Detection(目标检测)

• Rank & Sort Loss for Object Detection and Instance Segmentation
  :open_mouth:oral:star:code
• MDETR : Modulated Detection for End-to-End Multi-Modal Understanding
  :open_mouth:oral:star:code
• SimROD: A Simple Adaptation Method for Robust Object Detection
  :open_mouth:oral:house:project
  :newspaper:解读:ICCV2021 Oral SimROD：简单高效的数据增强！华为提出了一种简单的鲁棒目标检测自适应方法
• GraphFPN: Graph Feature Pyramid Network for Object Detection
• Fast Convergence of DETR with Spatially Modulated Co-Attention
  :star:code
• Oriented R-CNN for Object Detection
  :star:code
• Conditional DETR for Fast Training Convergence
  :newspaper:解读:通过显式寻找物体的 extremity 区域加快 DETR 的收敛：Conditional DETR
• Vector-Decomposed Disentanglement for Domain-Invariant Object Detection
  :star:code
• G-DetKD: Towards General Distillation Framework for Object Detectors via Contrastive and Semantic-guided Feature Imitation
• ODAM: Object Detection, Association, and Mapping using Posed RGB Video
  :open_mouth:oral
• Reconcile Prediction Consistency for Balanced Object Detection
• Deep Structured Instance Graph for Distilling Object Detectors
  :star:code
• Towards Rotation Invariance in Object Detection
  :star:code
• Morphable Detector for Object Detection on Demand
  :star:code
• DetCo: Unsupervised Contrastive Learning for Object Detection
  :star:code
• Domain-Invariant Disentangled Network for Generalizable Object Detection
• MDETR - Modulated Detection for End-to-End Multi-Modal Understanding
  :star:code
• Detecting Persuasive Atypicality by Modeling Contextual Compatibility
  :star:code
• Wanderlust: Online Continual Object Detection in the Real World
  :house:project
• PreDet: Large-Scale Weakly Supervised Pre-Training for Detection
• FMODetect: Robust Detection of Fast Moving Objects
• Multi-Source Domain Adaptation for Object Detection
• Self-Supervised Object Detection via Generative Image Synthesis
  :star:code
• Naturalistic Physical Adversarial Patch for Object Detectors
  :star:code
• Rethinking Transformer-Based Set Prediction for Object Detection
  :star:code
• Detecting Invisible People
  :house:project:tv:video
• Dynamic DETR: End-to-End Object Detection With Dynamic Attention
• CrossDet: Crossline Representation for Object Detection
  :star:code
• Robust Object Detection via Instance-Level Temporal Cycle Confusion
  :star:code
• End-to-End Semi-Supervised Object Detection With Soft Teacher
  :star:code
• Parallel Rectangle Flip Attack: A Query-Based Black-Box Attack Against Object Detection
• Fooling LiDAR Perception via Adversarial Trajectory Perturbation
  :star:code:house:project
• TOOD: Task-Aligned One-Stage Object Detection
  :open_mouth:oral:star:code
• Active Learning for Deep Object Detection via Probabilistic Modeling
  :star:code
  :newspaper:解读:ICCV2021 还在用大量数据暴力train模型？主动学习，教你选出数据集中最有价值的样本
• Dual Bipartite Graph Learning: A General Approach for Domain Adaptive Object Detection
• WB-DETR: Transformer-Based Detector without Backbone
• 3D目标检测
  • Geometry Uncertainty Projection Network for Monocular 3D Object Detection
    :star:code
  • Fog Simulation on Real LiDAR Point Clouds for 3D Object Detection in Adverse Weather
    :star:code:house:project
  • Is Pseudo-Lidar needed for Monocular 3D Object detection?
    :star:code
  • RandomRooms: Unsupervised Pre-training from Synthetic Shapes and Randomized Layouts for 3D Object Detection
  • LIGA-Stereo: Learning LiDAR Geometry Aware Representations for Stereo-based 3D Detector
    :star:code:house:project
  • Improving 3D Object Detection with Channel-wise Transformer
  • 4D-Net for Learned Multi-Modal Alignment
  • Pyramid R-CNN: Towards Better Performance and Adaptability for 3D Object Detection
  • Voxel Transformer for 3D Object Detection
  • An End-to-End Transformer Model for 3D Object Detection
    :open_mouth:oral:star:code:house:project
  • Unsupervised Domain Adaptive 3D Detection With Multi-Level Consistency
  • Group-Free 3D Object Detection via Transformers
    :star:code
  • VENet: Voting Enhancement Network for 3D Object Detection
  • Multi-Echo LiDAR for 3D Object Detection
  • Voxel Transformer for 3D Object Detection
  • RangeDet: In Defense of Range View for LiDAR-Based 3D Object Detection
    :star:code
  • The Devil Is in the Task: Exploiting Reciprocal Appearance-Localization Features for Monocular 3D Object Detection
  • Gated3D: Monocular 3D Object Detection From Temporal Illumination Cues
    :house:project
  • Are We Missing Confidence in Pseudo-LiDAR Methods for Monocular 3D Object Detection?
  • SPG: Unsupervised Domain Adaptation for 3D Object Detection via Semantic Point Generation
  • You Don't Only Look Once: Constructing Spatial-Temporal Memory for Integrated 3D Object Detection and Tracking
    :star:code:house:project:tv:video
  • Exploring Geometry-Aware Contrast and Clustering Harmonization for Self-Supervised 3D Object Detection
  • AutoShape: Real-Time Shape-Aware Monocular 3D Object Detection
    :star:code
  • Geometry-Based Distance Decomposition for Monocular 3D Object Detection
    :star:code
• 目标定位
  • Contrastive Attention Maps for Self-supervised Co-localization
  • TS-CAM: Token Semantic Coupled Attention Map for Weakly Supervised Object Localization
    :star:code
  • 弱监督目标定位
    • Normalization Matters in Weakly Supervised Object Localization
    • Online Refinement of Low-Level Feature Based Activation Map for Weakly Supervised Object Localization
    • Foreground Activation Maps for Weakly Supervised Object Localization
• Anomaly Detection(图像异常检测)
  • Divide-and-Assemble: Learning Block-wise Memory for Unsupervised Anomaly Detection
• 弱监督目标检测
  • Boosting Weakly Supervised Object Detection via Learning Bounding Box Adjusters
    :star:code
  • CaT: Weakly Supervised Object Detection With Category Transfer
    :star:code
• OOD 检测
  • Triggering Failures: Out-Of-Distribution detection by learning from local adversarial attacks in Semantic Segmentation
    :star:code
• 显著目标检测
  • Disentangled High Quality Salient Object Detection
  • Specificity-preserving RGB-D Saliency Detection
    :star:code
  • Light Field Saliency Detection with Dual Local Graph Learning and Reciprocative Guidance
  • MFNet: Multi-Filter Directive Network for Weakly Supervised Salient Object Detection
    :star:code
  • Scene Context-Aware Salient Object Detection
    :star:code
  • Dynamic Context-Sensitive Filtering Network for Video Salient Object Detection
    :star:code
  • iNAS: Integral NAS for Device-Aware Salient Object Detection
    :house:project
  • RGB-D显著目标检测
    • RGB-D Saliency Detection via Cascaded Mutual Information Minimization
      :star:code
  • co-saliency detection
    • Summarize and Search: Learning Consensus-aware Dynamic Convolution for Co-Saliency Detection
      :star:code
• 违禁物品检测
  • Towards Real-World Prohibited Item Detection: A Large-Scale X-ray Benchmark
    :sunflower:dataset
  • Towards Real-World X-Ray Security Inspection: A High-Quality Benchmark and Lateral Inhibition Module for Prohibited Items Detection
    :star:code
• 小样本目标检测
  • DeFRCN: Decoupled Faster R-CNN for Few-Shot Object Detection
    :star:code
  • Query Adaptive Few-Shot Object Detection With Heterogeneous Graph Convolutional Networks
  • Universal-Prototype Enhancing for Few-Shot Object Detection
    :star:code
• 视觉关系协同定位
  • Few-shot Visual Relationship Co-localization
    :star:code:house:project
• 密集目标检测
  • Mutual Supervision for Dense Object Detection
    :star:code
• 域适应目标检测
  • Seeking Similarities over Differences: Similarity-based Domain Alignment for Adaptive Object Detection
  • Knowledge Mining and Transferring for Domain Adaptive Object Detection
• 图像篡改检测
  • Image Manipulation Detection by Multi-View Multi-Scale Supervision
    :star:code
• Visual Relationship Detection(VRD视觉关系检测)
  • Grounding Consistency: Distilling Spatial Common Sense for Precise Visual Relationship Detection
• 长尾目标检测
  • Exploring Classification Equilibrium in Long-Tailed Object Detection
    :star:code
  • MosaicOS: A Simple and Effective Use of Object-Centric Images for Long-Tailed Object Detection
    :star:code
• Salient Object Ranking
  • Salient Object Ranking with Position-Preserved Attention
    :star:code
• 小目标检测
  • Robust Small Object Detection on the Water Surface Through Fusion of Camera and Millimeter Wave Radar
• 黑暗中目标检测
  • Multitask AET With Orthogonal Tangent Regularity for Dark Object Detection
    :star:code
• 3D object prediction
  • Holistic Pose Graph: Modeling Geometric Structure among Objects in a Scene using Graph Inference for 3D Object Prediction
    :star:code
• 多目标检测
  • Training Multi-Object Detector by Estimating Bounding Box Distribution for Input Image
    :star:code
• 3D object grounding
  • Free-form Description Guided 3D Visual Graph Network for Object Grounding in Point Cloud
    :star:code
• 细粒度裂纹检测
  • CrackFormer: Transformer Network for Fine-Grained Crack Detection
• 线段检测
  • ELSD: Efficient Line Segment Detector and Descriptor
• 细胞检测与分类
  • Multi-Class Cell Detection Using Spatial Context Representation
    :star:code
• 阴影检测
  • Mitigating Intensity Bias in Shadow Detection via Feature Decomposition and Reweighting
• 社交距离检测
  • BEV-Net: Assessing Social Distancing Compliance by Joint People Localization and Geometric Reasoning
    :star:code
• 伪装目标检测
  • Uncertainty-Guided Transformer Reasoning for Camouflaged Object Detection
    :star:code

13.Image Segmentation(图像分割)

• Standardized Max Logits: A Simple yet Effective Approach for Identifying Unexpected Road Obstacles in Urban-Scene Segmentation
  :open_mouth:oral:star:code:tv:video
• TransForensics: Image Forgery Localization with Dense Self-Attention
• From Contexts to Locality: Ultra-high Resolution Image Segmentation via Locality-aware Contextual Correlation
  :star:code
• Labels4Free: Unsupervised Segmentation using StyleGAN
  :house:project:tv:video
• Warp-Refine Propagation: Semi-Supervised Auto-labeling via Cycle-consistency
• Scaling up instance annotation via label propagation
  :star:code:house:project
• Robust Trust Region for Weakly Supervised Segmentation
  :star:code:tv:video
• HPNet: Deep Primitive Segmentation Using Hybrid Representations
  :star:code
• Weakly Supervised Segmentation of Small Buildings With Point Labels
• BAPA-Net: Boundary Adaptation and Prototype Alignment for Cross-Domain Semantic Segmentation
  :star:code
• Conditional Diffusion for Interactive Segmentation
• Human Detection and Segmentation via Multi-view Consensus
  :star:code
• Unidentified Video Objects: A Benchmark for Dense, Open-World Segmentation
• Enhanced Boundary Learning for Glass-Like Object Segmentation
  :star:code
• PARTS: Unsupervised segmentation with slots, attention and independence maximization
• Predictive Feature Learning for Future Segmentation Prediction
• Perception-Aware Multi-Sensor Fusion for 3D LiDAR Semantic Segmentation
  :star:code
• Segmenter: Transformer for Semantic Segmentation
  :star:code
• C3-SemiSeg: Contrastive Semi-Supervised Segmentation via Cross-Set Learning and Dynamic Class-Balancing
• 全景分割
  • Panoptic Narrative Grounding
• 语义分割
  • Re-distributing Biased Pseudo Labels for Semi-supervised Semantic Segmentation: A Baseline Investigation
    :open_mouth:oral:star:code
  • Personalized Image Semantic Segmentation
    :star:code
  • RECALL: Replay-based Continual Learning in Semantic Segmentation
    :star:code
  • Deep Metric Learning for Open World Semantic Segmentation
  • LabOR: Labeling Only if Required for Domain Adaptive Semantic Segmentation
    :open_mouth:oral
  • Dual Path Learning for Domain Adaptation of Semantic Segmentation
    :star:code
  • Multi-Target Adversarial Frameworks for Domain Adaptation in Semantic Segmentation
  • Exploiting a Joint Embedding Space for Generalized Zero-Shot Semantic Segmentation
    :star:code:house:project
  • Multi-Anchor Active Domain Adaptation for Semantic Segmentation
    :open_mouth:oral:star:code
  • Pixel Contrastive-Consistent Semi-Supervised Semantic Segmentation
  • Self-Regulation for Semantic Segmentation
    :star:code
  • ShapeConv: Shape-aware Convolutional Layer for Indoor RGB-D Semantic Segmentation
    :star:code
  • Generalize then Adapt: Source-Free Domain Adaptive Semantic Segmentation
    :house:project
  • Mining Contextual Information Beyond Image for Semantic Segmentation
    :star:code
  • ISNet: Integrate Image-Level and Semantic-Level Context for Semantic Segmentation
    :star:code
  • Pseudo-mask Matters in Weakly-supervised Semantic Segmentation
    :star:code
  • SIGN: Spatial-information Incorporated Generative Network for Generalized Zero-shot Semantic Segmentation
  • Region-Aware Contrastive Learning for Semantic Segmentation
  • GP-S3Net: Graph-Based Panoptic Sparse Semantic Segmentation Network
  • Domain Adaptive Semantic Segmentation With Self-Supervised Depth Estimation
    :star:code
  • Scribble-Supervised Semantic Segmentation by Uncertainty Reduction on Neural Representation and Self-Supervision on Neural Eigenspace
  • Exploring Cross-Image Pixel Contrast for Semantic Segmentation
    :open_mouth:oral:star:code
  • Dynamic Divide-and-Conquer Adversarial Training for Robust Semantic Segmentation
    :star:code
  • Uncertainty-Aware Pseudo Label Refinery for Domain Adaptive Semantic Segmentation
  • Contrastive Learning for Label Efficient Semantic Segmentation
  • Scaling Semantic Segmentation Beyond 1K Classes on a Single GPU
    :star:code
  • Prototypical Matching and Open Set Rejection for Zero-Shot Semantic Segmentation
  • Geometric Unsupervised Domain Adaptation for Semantic Segmentation
  • Calibrated Adversarial Refinement for Stochastic Semantic Segmentation
    :star:code
  • Multi-View Radar Semantic Segmentation
    :star:code
  • Exploring Robustness of Unsupervised Domain Adaptation in Semantic Segmentation
    :open_mouth:oral:star:code
  • Specialize and Fuse: Pyramidal Output Representation for Semantic Segmentation
  • Unsupervised Semantic Segmentation by Contrasting Object Mask Proposals
    :star:code
  • Scribble-Supervised Semantic Segmentation Inference
  • Semi-Supervised Semantic Segmentation With Pixel-Level Contrastive Learning From a Class-Wise Memory Bank
    :star:code
  • 小样本语义分割
    • Simpler is Better: Few-shot Semantic Segmentation with Classifier Weight Transformer
      :star:code
    • Learning Meta-class Memory for Few-Shot Semantic Segmentation
    • Few-Shot Semantic Segmentation With Cyclic Memory Network
  • 3D语义分割
    • VMNet: Voxel-Mesh Network for Geodesic-Aware 3D Semantic Segmentation
      :open_mouth:oral:star:code
    • Sparse-to-Dense Feature Matching: Intra and Inter Domain Cross-Modal Learning in Domain Adaptation for 3D Semantic Segmentation
      :star:code
    • Weakly Supervised 3D Semantic Segmentation Using Cross-Image Consensus and Inter-Voxel Affinity Relations
  • 视频语义分割
    • Domain Adaptive Video Segmentation via Temporal Consistency Regularization
      :star:code
  • 弱监督语义分割
    • Leveraging Auxiliary Tasks with Affinity Learning for Weakly Supervised Semantic Segmentation
      :star:code
    • Complementary Patch for Weakly Supervised Semantic Segmentation
    • ECS-Net: Improving Weakly Supervised Semantic Segmentation by Using Connections Between Class Activation Maps
    • Unlocking the Potential of Ordinary Classifier: Class-Specific Adversarial Erasing Framework for Weakly Supervised Semantic Segmentation
      :star:code
    • Context Decoupling Augmentation for Weakly Supervised Semantic Segmentation
      :star:code
    • Seminar Learning for Click-Level Weakly Supervised Semantic Segmentation
  • 点云语义分割
    • ReDAL: Region-based and Diversity-aware Active Learning for Point Cloud Semantic Segmentation
      :tv:video
    • Perturbed Self-Distillation: Weakly Supervised Large-Scale Point Cloud Semantic Segmentation
    • TempNet: Online Semantic Segmentation on Large-Scale Point Cloud Series
    • Guided Point Contrastive Learning for Semi-Supervised Point Cloud Semantic Segmentation
    • Learning With Noisy Labels for Robust Point Cloud Segmentation
      :star:code:house:project
  • OOD
    • Entropy Maximization and Meta Classification for Out-of-Distribution Detection in Semantic Segmentation
      :star:code
• 实例分割
  • Rank & Sort Loss for Object Detection and Instance Segmentation
    :open_mouth:oral:star:code
  • SOTR: Segmenting Objects with Transformers
    :star:code
  • A Weakly Supervised Amodal Segmenter with Boundary Uncertainty Estimation
  • Instances as Queries
    :star:code:tv:video
  • CrossVIS: Crossover Learning for Fast Online Video Instance Segmentation
    :star:code:tv:video
  • CDNet: Centripetal Direction Network for Nuclear Instance Segmentation
    :star:code
  • PrimitiveNet: Primitive Instance Segmentation With Local Primitive Embedding Under Adversarial Metric
    :star:code
  • FASA: Feature Augmentation and Sampling Adaptation for Long-Tailed Instance Segmentation
    :star:code:house:project
  • Prior to Segment: Foreground Cues for Weakly Annotated Classes in Partially Supervised Instance Segmentation
    :star:code
  • DiscoBox: Weakly Supervised Instance Segmentation and Semantic Correspondence From Box Supervision
  • End-to-End Video Instance Segmentation via Spatial-Temporal Graph Neural Networks
  • The Surprising Impact of Mask-Head Architecture on Novel Class Segmentation
    :house:project
  • How Shift Equivariance Impacts Metric Learning for Instance Segmentation
    :star:code
  • Parallel Detection-and-Segmentation Learning for Weakly Supervised Instance Segmentation
  • Real-Time Instance Segmentation With Discriminative Orientation Maps
    :star:code
  • 视频实例分割
    • Video Instance Segmentation with a Propose-Reduce Paradigm
      :star:code
  • 3D实例分割
    • Hierarchical Aggregation for 3D Instance Segmentation
      :star:code
• 小样本分割
  • Mining Latent Classes for Few-shot Segmentation
    :open_mouth:oral:star:code
• Human Motion Segmentation(人体运动分割)
  • Graph Constrained Data Representation Learning for Human Motion Segmentation
    :star:code
  • Hypercorrelation Squeeze for Few-Shot Segmenation
    :star:code:house:project
• 点云分割
  • Learning with Noisy Labels for Robust Point Cloud Segmentation
    :open_mouth:oral:star:code:house:project
  • DRINet: A Dual-Representation Iterative Learning Network for Point Cloud Segmentation
  • RPVNet: A Deep and Efficient Range-Point-Voxel Fusion Network for LiDAR Point Cloud Segmentation
• 视频目标分割(VOS)
  • Full-Duplex Strategy for Video Object Segmentation
    :house:project
  • Joint Inductive and Transductive Learning for Video Object Segmentation
    :star:code
  • Hierarchical Memory Matching Network for Video Object Segmentation
    :star:code
  • Self-supervised Video Object Segmentation by Motion Grouping
    :star:code:house:project:tv:video
  • Deep Transport Network for Unsupervised Video Object Segmentation
  • Generating Masks From Boxes by Mining Spatio-Temporal Consistencies in Videos
    :star:code
  • Learning Motion-Appearance Co-Attention for Zero-Shot Video Object Segmentation
    :star:code
  • Video Object Segmentation With Dynamic Memory Networks and Adaptive Object Alignment
    :star:code
• 语义场景分割
  • BiMaL: Bijective Maximum Likelihood Approach to Domain Adaptation in Semantic Scene Segmentation
    :star:code
• Referring Segmentation(基于文本的分割)
  • Vision-Language Transformer and Query Generation for Referring Segmentation
    :star:code
• 场景理解
  • DeepPanoContext: Panoramic 3D Scene Understanding with Holistic Scene Context Graph and Relation-based Optimization
    :open_mouth:oral:star:code:house:project:tv:video
  • ACDC: The Adverse Conditions Dataset With Correspondences for Semantic Driving Scene Understanding
    :house:project
  • Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understanding
    :star:code
• CMA
  • Towards Better Explanations of Class Activation Mapping
  • Keep CALM and Improve Visual Feature Attribution
    :star:code
  • LFI-CAM: Learning Feature Importance for Better Visual Explanation
    :star:code
• 多目标分割
  • [Faster Multi-Object Segmentation Using Parallel Quadratic Pseudo-Boolean Optimization](https://openaccess.thecvf.com/content/ICCV2021/papers/Jeppesen_