CvHub

深度解讀電腦視覺論文 -- 經典到前沿 -- 中英雙語對照 + 多維度深度解析

經典論文 (Classics)

論文	領域
Deep Residual Learning for Image Recognition (ResNet)	影像分類 (CVPR 2016 Best Paper)
You Only Look Once (YOLO)	物件偵測 (CVPR 2016)
Attention Is All You Need (Transformer)	序列模型 (NeurIPS 2017)
An Image is Worth 16x16 Words (ViT)	Vision Transformer (ICLR 2021)
Denoising Diffusion Probabilistic Models (DDPM)	擴散模型 (NeurIPS 2020)

CVPR (2013-2025, 13 屆, 156 篇)

含經典論文 5 篇 + CVPR 156 篇 + ICCV 84 篇 + ECCV 72 篇 = 共 317 篇

CVPR 2013

論文	領域
Fast, Accurate Detection of 100,000 Object Classes on a Single Machine	物件偵測 (Best Paper)
Discriminative Non-blind Deblurring	影像復原 (Best Student Paper)
Lost! Leveraging the Crowd for Probabilistic Visual Self-Localization	視覺定位 (Honorable Mention)
Online Object Tracking: A Benchmark (OTB)	目標追蹤
SLAM++: Simultaneous Localisation and Mapping at the Level of Objects	SLAM
Pedestrian Detection with Unsupervised Multi-Stage Feature Learning	行人偵測
Perceptual Organization and Recognition of Indoor Scenes from RGB-D Images	RGB-D 場景理解
HON4D: Histogram of Oriented 4D Normals for Activity Recognition	動作辨識
Joint 3D Scene Reconstruction and Class Segmentation	3D 重建
Hierarchical Saliency Detection	顯著性偵測
Detecting and Aligning Faces by Image Retrieval	人臉偵測
Multi-target Tracking by Lagrangian Relaxation to Min-cost Network Flow	多目標追蹤

CVPR 2014

論文	領域
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation (R-CNN)	物件偵測
DeepFace: Closing the Gap to Human-Level Performance in Face Verification	人臉辨識
What Camera Motion Reveals About Shape with Unknown BRDF	3D 形狀 (Best Paper)
Learning and Transferring Mid-Level Image Representations using CNNs	遷移學習
Multiscale Combinatorial Grouping (MCG)	分割
Partial Optimality by Pruning for MAP-inference	最佳化 (Best Student Paper)
3D Shape and Indirect Appearance by Structured Light Transport	3D 光傳輸 (Honorable Mention)
Deep Learning Face Representation from Predicting 10,000 Classes (DeepID)	人臉辨識
Caffe: Convolutional Architecture for Fast Feature Embedding	框架
Human Action Recognition by Representing 3D Skeletons as Points in a Lie Group	動作辨識
Fast and Accurate Image Matching with Cascade Hashing for 3D Reconstruction	3D 重建
Multi-Object Tracking via Constrained Sequential Labeling	多目標追蹤

CVPR 2015

論文	領域
DynamicFusion: Reconstruction and Tracking of Non-rigid Scenes in Real-Time	3D 重建 (Best Paper)
Fully Convolutional Networks for Semantic Segmentation (FCN)	語義分割 (Honorable Mention)
Going Deeper with Convolutions (GoogLeNet/Inception)	網路架構
FaceNet: A Unified Embedding for Face Recognition and Clustering	人臉辨識
Show and Tell: A Neural Image Caption Generator	影像描述
Long-Term Recurrent Convolutional Networks (LRCN)	影片理解
Hypercolumns for Object Segmentation and Fine-grained Localization	分割
3D ShapeNets: A Deep Representation for Volumetric Shapes	3D 形狀
Category-Specific Object Reconstruction from a Single Image	3D 重建 (Best Student Paper)
Picture: A Probabilistic Programming Language for Scene Perception	場景理解 (Honorable Mention)
Deformable Part Models are Convolutional Neural Networks	物件偵測
Hierarchical Recurrent Neural Network for Skeleton Based Action Recognition	動作辨識

CVPR 2016

論文	領域
Structural-RNN: Deep Learning on Spatio-Temporal Graphs	時空圖 (Best Student Paper)
Sublabel-Accurate Relaxation of Nonconvex Energies	最佳化 (Honorable Mention)
Rethinking the Inception Architecture for Computer Vision (Inception v3)	網路架構
Learning Deep Features for Discriminative Localization (CAM)	可解釋性
Context Encoders: Feature Learning by Inpainting	影像修復
Deep Sliding Shapes for Amodal 3D Object Detection in RGB-D Images	3D 偵測
Stacked Attention Networks for Image Question Answering	VQA
NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis	動作辨識
Training Region-based Object Detectors with Online Hard Example Mining (OHEM)	物件偵測
Temporal Action Localization in Untrimmed Videos via Multi-stage CNNs	影片理解
Learning Dense Correspondence via 3D-Guided Cycle Consistency	3D 對應
Single Image 3D Interpreter Network	3D 理解

CVPR 2017

論文	領域
Densely Connected Convolutional Networks (DenseNet)	網路架構 (Best Paper)
Learning from Simulated and Unsupervised Images through Adversarial Training (SimGAN)	GAN (Best Paper)
PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation	3D 點雲
Feature Pyramid Networks for Object Detection (FPN)	物件偵測
Image-to-Image Translation with Conditional Adversarial Networks (pix2pix)	GAN
Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields (OpenPose)	姿態估計
Pyramid Scene Parsing Network (PSPNet)	語義分割
Aggregated Residual Transformations for Deep Neural Networks (ResNeXt)	網路架構
Photo-Realistic Single Image Super-Resolution Using a GAN (SRGAN)	超解析度
Deformable Convolutional Networks	物件偵測
Annotating Object Instances with a Polygon-RNN	實例分割 (Honorable Mention)
Computational Imaging on the Electric Grid	計算攝影 (Best Student Paper)

CVPR 2018

論文	領域
Taskonomy: Disentangling Task Transfer Learning	遷移學習 (Best Paper)
Total Capture: A 3D Deformation Model for Tracking Faces, Hands, and Bodies	3D 人體 (Best Student Paper)
Non-local Neural Networks	影片理解
Squeeze-and-Excitation Networks (SENet)	網路架構
StarGAN: Unified GANs for Multi-Domain Image-to-Image Translation	GAN
MobileNetV2: Inverted Residuals and Linear Bottlenecks	高效架構
High-Resolution Image Synthesis with Conditional GANs (pix2pixHD)	影像合成
DensePose: Dense Human Pose Estimation In The Wild	姿態估計
Progressive Growing of GANs (ProGAN)	GAN
A Closer Look at Spatiotemporal Convolutions for Action Recognition (R(2+1)D)	影片理解
Deep Learning of Graph Matching	圖匹配 (Honorable Mention)
SPLATNet: Sparse Lattice Networks for Point Cloud Processing	點雲 (Honorable Mention)

CVPR 2019

論文	領域
A Theory of Fermat Paths for Non-Line-of-Sight Shape Reconstruction	非視線重建 (Best Paper)
ArcFace: Additive Angular Margin Loss for Deep Face Recognition	人臉辨識
A Style-Based Generator Architecture for GANs (StyleGAN)	GAN / 影像生成
Semantic Image Synthesis with Spatially-Adaptive Normalization (SPADE)	語義影像合成
DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation	3D 形狀表示
PointRCNN: 3D Object Proposal Generation and Detection From Point Cloud	3D 物件偵測
Fast Online Object Tracking and Segmentation (SiamMask)	物件追蹤 / 分割
Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentation	神經架構搜尋
Libra R-CNN: Towards Balanced Learning for Object Detection	物件偵測
Bag of Tricks for Image Classification with CNNs	影像分類
Generalized Intersection over Union (GIoU)	物件偵測
GANFIT: GAN Fitting for High Fidelity 3D Face Reconstruction	3D 人臉重建

CVPR 2020

論文	領域
Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild	3D 重建 (Best Paper)
BSP-Net: Generating Compact Meshes via Binary Space Partitioning	3D 形狀生成 (Best Student Paper)
Momentum Contrast for Unsupervised Visual Representation Learning (MoCo)	自監督學習
EfficientDet: Scalable and Efficient Object Detection	物件偵測
PointRend: Image Segmentation as Rendering	影像分割
3D Photography using Context-aware Layered Depth Inpainting	3D 攝影
HigherHRNet: Scale-Aware Representation Learning for Bottom-Up Human Pose Estimation	人體姿態估計
PIFuHD: Multi-Level Pixel-Aligned Implicit Function for High-Resolution 3D Human Digitization	3D 人體重建
Circle Loss: A Unified Perspective of Pair Similarity Optimization	度量學習
X3D: Expanding Architectures for Efficient Video Recognition	影片辨識
DeepCap: Monocular Human Performance Capture Using Weak Supervision	人體動態捕捉
Adversarial Latent Autoencoders (ALAE)	生成模型

CVPR 2021

論文	領域
GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields	場景生成 (Best Paper)
Exploring Simple Siamese Representation Learning (SimSiam)	自監督學習 (Honorable Mention)
NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections	新視角合成
RepVGG: Making VGG-Style ConvNets Great Again	網路架構
Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective (SETR)	語義分割
D-NeRF: Neural Radiance Fields for Dynamic Scenes	動態場景
Real-Time High-Resolution Background Matting	影像摳圖 (Honorable Mention)
End-to-End Video Instance Segmentation with Transformers (VisTR)	影片實例分割
RAFT-3D: Scene Flow Using Rigid-Motion Embeddings	場景流
Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet	Vision Transformer
UP-DETR: Unsupervised Pre-training for Object Detection with Transformers	物件偵測
Learning High Fidelity Depths of Dressed Humans by Watching Social Media Dance Videos	人體深度估計 (Honorable Mention)

CVPR 2022

論文	領域
Learning to Solve Hard Minimal Problems	幾何視覺 (Best Paper)
Dual-Shutter Optical Vibration Sensing	計算攝影 (Best Paper Honorable Mention)
EPro-PnP: Generalized End-to-End Probabilistic Perspective-n-Points for Monocular Object Pose Estimation	姿態估計 (Best Student Paper)
Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields	神經輻射場 (Best Student Paper Honorable Mention)
Masked Autoencoders Are Scalable Vision Learners (MAE)	自監督學習
High-Resolution Image Synthesis with Latent Diffusion Models (LDM)	擴散模型
A ConvNet for the 2020s (ConvNeXt)	網路架構
Swin Transformer V2: Scaling Up Capacity and Resolution	Vision Transformer
Masked-attention Mask Transformer for Universal Image Segmentation (Mask2Former)	統一分割
DN-DETR: Accelerate DETR Training by Introducing Query DeNoising	物體偵測
Restormer: Efficient Transformer for High-Resolution Image Restoration	影像修復
Point-NeRF: Point-based Neural Radiance Fields	3D 重建

CVPR 2023

論文	領域
Visual Programming: Compositional Visual Reasoning Without Training	視覺推理 (Best Paper)
Planning-oriented Autonomous Driving (UniAD)	自動駕駛 (Best Paper)
DynIBaR: Neural Dynamic Image-Based Rendering	新視角合成 (Best Paper Honorable Mention)
3D Registration with Maximal Cliques	點雲配準 (Best Student Paper)
DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation	擴散模型 (Best Student Paper Honorable Mention)
InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutions	視覺基礎模型
ImageBind: One Embedding Space To Bind Them All	多模態學習
EVA: Exploring the Limits of Masked Visual Representation Learning at Scale	自監督學習
ODISE: Open-Vocabulary Panoptic Segmentation with Text-to-Image Diffusion Models	開放詞彙分割
Scaling Up GANs for Text-to-Image Synthesis (GigaGAN)	GAN / 影像生成
Mask DINO: Towards A Unified Transformer-based Framework for Object Detection and Segmentation	偵測與分割
Score Jacobian Chaining: Lifting Pretrained 2D Diffusion Models for 3D Generation	3D 生成

CVPR 2024

論文	領域
Generative Image Dynamics	影像動態生成 (Best Paper)
Rich Human Feedback for Text-to-Image Generation	文字到影像 (Best Paper)
Mip-Splatting: Alias-free 3D Gaussian Splatting	3D 高斯潑灑 (Best Student Paper)
BioCLIP: A Vision Foundation Model for the Tree of Life	生物視覺基礎模型 (Best Student Paper)
pixelSplat: 3D Gaussian Splats from Image Pairs	3D 重建 (Honorable Mention)
Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data	深度估計
YOLO-World: Real-Time Open-Vocabulary Object Detection	開放詞彙偵測
Florence-2: Advancing a Unified Representation for a Variety of Vision Tasks	多模態基礎模型
InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks	視覺語言模型
4D Gaussian Splatting for Real-Time Dynamic Scene Rendering	動態 3D 渲染
EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anything	高效分割
Objects as Volumes: A Stochastic Geometry View of Opaque Solids	渲染理論 (Honorable Mention)

CVPR 2025

論文	領域
VGGT: Visual Geometry Grounded Transformer	3D 多視角幾何 (Best Paper)
Neural Inverse Rendering from Propagating Light	逆渲染 (Best Student Paper)
MegaSaM: Accurate, Fast and Robust Structure and Motion from Casual Dynamic Videos	結構與運動 (Honorable Mention)
Navigation World Models	世界模型 (Honorable Mention)
Molmo and PixMo: Open Weights and Open Data for State-of-the-Art VLMs	視覺語言模型 (Honorable Mention)
3D Student Splatting and Scooping	神經渲染 (Honorable Mention)
Generative Multimodal Pretraining with Discrete Diffusion Timestep Tokens (DDT-LLaMA)	多模態預訓練 (Best Student Paper Honorable Mention)
MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors	即時 SLAM
Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation	統一多模態
TRELLIS: Structured 3D Latents for Scalable and Versatile 3D Generation	3D 生成
DepthCrafter: Generating Consistent Long Depth Sequences for Open-world Videos	影片深度估測
OmniGen: Unified Image Generation	統一影像生成

ICCV (2013-2025, 7 屆, 84 篇)

ICCV 2013

論文	領域
From Large Scale Image Categorization to Entry-Level Categories	影像分類 (Marr Prize)
Hierarchical Data-driven Descent for Efficient Optimal Deformation Estimation	形變估計 (Honorable Mention)
Piecewise Rigid Scene Flow	3D 場景流 (Honorable Mention)
Action Recognition with Improved Trajectories	動作辨識
Structured Forests for Fast Edge Detection	邊緣偵測
OverFeat: Integrated Recognition, Localization and Detection	物件偵測
Saliency Detection: A Boolean Map Approach	顯著偵測
Holistic Scene Understanding for 3D Object Detection with RGBD	3D 偵測
Saliency Detection via Absorbing Markov Chain	顯著偵測
Segmentation Driven Object Detection with Fisher Vectors	物件偵測
PhotoOCR: Reading Text in Uncontrolled Conditions	文字辨識
Robust Object Tracking with Online Multi-lifespan Dictionary Learning	視覺追蹤

ICCV 2015

論文	領域
Deep Neural Decision Forests	分類 (Marr Prize)
Holistically-Nested Edge Detection (HED)	邊緣偵測 (Honorable Mention)
Fast R-CNN	物件偵測
Delving Deep into Rectifiers (PReLU)	影像分類
FlowNet: Learning Optical Flow with Convolutional Networks	光流估計
Learning Spatiotemporal Features with 3D Convolutional Networks (C3D)	影片理解
Conditional Random Fields as Recurrent Neural Networks (CRF-RNN)	語意分割
Learning Deconvolution Network for Semantic Segmentation	語意分割
Unsupervised Visual Representation Learning by Context Prediction	自監督學習
Ask Your Neurons: A Neural-Based Approach to Answering Questions	視覺問答
Unsupervised Learning of Visual Representations using Videos	自監督學習
Dense Optical Flow Prediction From a Static Image	運動預測

ICCV 2017

論文	領域
Mask R-CNN	實例分割 (Marr Prize)
Focal Loss for Dense Object Detection (RetinaNet)	物件偵測 (Best Student Paper)
First-Person Activity Forecasting	活動預測 (Honorable Mention)
CycleGAN: Unpaired Image-to-Image Translation	影像生成
Grad-CAM: Visual Explanations from Deep Networks	可解釋性
Channel Pruning for Accelerating Very Deep Neural Networks	模型壓縮
Open Set Domain Adaptation	領域適應
Globally-Optimal Inlier Set Maximisation	3D 視覺
Structured Attentions for Visual Question Answering	VQA
Generative Image Inpainting with Contextual Attention	影像修補
Globally and Locally Consistent Image Completion	影像補全
Learning to Segment Every Thing	實例分割

ICCV 2019

論文	領域
SinGAN: Learning a Generative Model from a Single Natural Image	影像生成 (Marr Prize)
PLMP: Point-Line Minimal Problems	多視圖幾何 (Best Student Paper)
CutMix: Regularization Strategy to Train Strong Classifiers	資料增強
FCOS: Fully Convolutional One-Stage Object Detection	物件偵測
Deep Hough Voting for 3D Object Detection (VoteNet)	3D 偵測
SlowFast Networks for Video Recognition	影片理解
Mesh R-CNN	3D 重建
Larger Norm More Transferable	領域適應
Exploring Randomly Wired Neural Networks	架構搜尋
Cascade R-CNN	物件偵測
Asynchronous Single-Photon 3D Imaging	計算攝影 (Honorable Mention)
Specifying Object Attributes and Relations in Interactive Scene Generation	場景生成 (Honorable Mention)

ICCV 2021

論文	領域
Swin Transformer	Vision Transformer (Marr Prize)
Pixel-Perfect Structure-from-Motion	SfM (Best Student Paper)
Mip-NeRF: Anti-Aliasing Neural Radiance Fields	神經輻射場 (Honorable Mention)
OpenGAN: Open-Set Recognition via Open Data Generation	開放集辨識 (Honorable Mention)
Common Objects in 3D (CO3D)	3D 資料集 (Honorable Mention)
Vision Transformers for Dense Prediction (DPT)	稠密預測
Pyramid Vision Transformer (PVT)	Vision Transformer
Focal Transformer	Vision Transformer
NeRF in the Dark	神經輻射場
MaskFormer: Per-Pixel Classification is Not All You Need	語意分割
Viewing Graph Solvability via Cycle Consistency	多視圖幾何 (Honorable Mention)
Multiscale Vision Transformers (MViT)	Vision Transformer

ICCV 2023

論文	領域
Passive Ultra-Wideband Single-Photon Imaging	計算攝影 (Marr Prize)
ControlNet: Adding Conditional Control to Text-to-Image Diffusion Models	可控生成 (Marr Prize)
Segment Anything (SAM)	分割基礎模型 (Honorable Mention)
Tracking Everything Everywhere All at Once	運動追蹤 (Best Student Paper)
3D Gaussian Splatting for Real-Time Radiance Field Rendering	3D 渲染
DINOv2: Learning Robust Visual Features without Supervision	自監督學習
AnyDoor: Zero-shot Object-level Image Customization	影像編輯
LAVIE: High-Quality Video Generation	影片生成
Nerfstudio: A Modular Framework for Neural Radiance Fields	神經輻射場
LERF: Language Embedded Radiance Fields	3D 語言
IP-Adapter: Text Compatible Image Prompt Adapter	可控生成
Tracking Anything in High Quality	追蹤

ICCV 2025

論文	領域
BrickGPT: Generating Physically Stable Brick Structures from Text	3D 生成 (Marr Prize)
Spatially-Varying Autofocus	計算攝影 (Honorable Mention)
RayZer: A Self-supervised Large View Synthesis Model	新視角合成 (Best Student Paper HM)
FlowEdit: Inversion-Free Text-Based Editing	影像編輯
LLaVA-CoT: Let Vision Language Models Reason Step-by-Step	視覺語言模型
Sa2VA: Marrying SAM2 with LLaVA	多模態定位
Dynamic-DINO: MoE for Real-time Open-Vocabulary Detection	開放詞彙偵測
MaskControl: Spatio-Temporal Control for Motion Synthesis	動作生成
HERMES: A Unified Self-Driving World Model	自動駕駛
LongSplat: Robust Unposed 3D Gaussian Splatting	3D 重建
SceneSplat: Gaussian Splatting-based Scene Understanding	3D 場景理解
From Image to Video: An Empirical Study of Diffusion Representations	表示學習

ECCV (2014-2024, 6 屆, 72 篇)

ECCV 2014

論文	領域
Large-Scale Object Classification using Label Relation Graphs	物件分類 (Best Paper)
Scene Chronology	場景年代學 (Best Paper)
Microsoft COCO: Common Objects in Context	資料集 / 物件偵測
Spatial Pyramid Pooling in Deep Convolutional Networks (SPPNet)	影像分類 / 偵測
Visualizing and Understanding Convolutional Networks (ZFNet)	網路視覺化
Image Super-Resolution Using Deep Convolutional Networks (SRCNN)	超解析度
LSD-SLAM: Large-Scale Direct Monocular SLAM	SLAM
CNN Features Off-the-Shelf: An Astounding Baseline	遷移學習
Edge Boxes: Locating Object Proposals from Edges	物件候選框
Simultaneous Detection and Segmentation	偵測與分割
Part-Based R-CNNs for Fine-Grained Category Detection	細粒度偵測
Action Recognition with Stacked Fisher Vectors	動作辨識

ECCV 2016

論文	領域
Real-Time 3D Reconstruction and 6-DoF Tracking with an Event Camera	事件相機 (Best Paper)
The Fast Bilateral Solver	影像處理 (Honorable Mention)
SSD: Single Shot MultiBox Detector	物件偵測
Identity Mappings in Deep Residual Networks	網路架構
Perceptual Losses for Real-Time Style Transfer and Super-Resolution	風格轉換 / 超解析
Colorful Image Colorization	影像上色
Fully-Convolutional Siamese Networks for Object Tracking (SiamFC)	視覺追蹤
Stacked Hourglass Networks for Human Pose Estimation	姿態估測
Learning to Track at 100 FPS with Deep Regression Networks (GOTURN)	視覺追蹤
Wide Residual Networks (WRN)	網路架構
ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation	語義分割
Temporal Segment Networks (TSN)	影片動作辨識

ECCV 2018

論文	領域
Implicit 3D Orientation Learning for 6D Object Detection	6D 偵測 (Best Paper)
Group Normalization	正規化 (Honorable Mention)
GANimation: Anatomically-aware Facial Animation from a Single Image	臉部動畫 (Honorable Mention)
CBAM: Convolutional Block Attention Module	注意力機制
CornerNet: Detecting Objects as Paired Keypoints	物件偵測
Encoder-Decoder with Atrous Separable Convolution (DeepLabv3+)	語義分割
Simple Baselines for Human Pose Estimation and Tracking	姿態估測
ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design	高效網路
PersonLab: Person Pose Estimation and Instance Segmentation	人體分析
Exploring the Limits of Weakly Supervised Pretraining	弱監督預訓練
ESPNet: Efficient Spatial Pyramid of Dilated Convolutions	即時語義分割
Rethinking ImageNet Pre-training	預訓練策略

ECCV 2020

論文	領域
RAFT: Recurrent All-Pairs Field Transforms for Optical Flow	光流估測 (Best Paper)
NeRF: Representing Scenes as Neural Radiance Fields	神經渲染 (Honorable Mention)
Towards Streaming Perception	串流感知 (Honorable Mention)
DETR: End-to-End Object Detection with Transformers	物件偵測
SOLO: Segmenting Objects by Locations	實例分割
Object-Contextual Representations for Semantic Segmentation (OCRNet)	語義分割
Rewriting a Deep Generative Model	生成模型編輯
PointContrast: Unsupervised Pre-training for 3D Point Cloud Understanding	3D 點雲
Rethinking Bottleneck Structure for Efficient Mobile Network Design	行動網路
AR-Net: Adaptive Frame Resolution for Efficient Action Recognition	動作辨識
Knowledge Distillation Meets Self-Supervision	知識蒸餾
Unpaired Learning of Deep Image Denoising	影像去噪

ECCV 2022

論文	領域
On the Versatile Uses of Partial Distance Correlation in Deep Learning	深度學習理論 (Best Paper)
Level Set Theory for Neural Implicit Evolution	神經隱式表面 (Honorable Mention)
Pose-NDF: Modelling Human Pose Manifolds with Neural Distance Fields	人體姿態 (Honorable Mention)
BEVFormer: Learning Bird's-Eye-View Representation	自動駕駛
TensoRF: Tensorial Radiance Fields	神經渲染
MaxViT: Multi-Axis Vision Transformer	視覺 Transformer
DaViT: Dual Attention Vision Transformers	視覺 Transformer
Panoptic Scene Graph Generation	場景圖生成
PETR: Position Embedding Transformation for Multi-View 3D Detection	3D 偵測
2DPASS: 2D Priors Assisted Semantic Segmentation on LiDAR	點雲分割
AnyDoor: Zero-shot Object-level Image Customization	影像編輯
MaskGIT: Masked Generative Image Transformer	影像生成

ECCV 2024

論文	領域
Minimalist Vision with Freeform Pixels	計算攝影 (Best Paper)
Rasterized Edge Gradients: Handling Discontinuities Differentiably	可微分渲染 (Honorable Mention)
Concept Arithmetics for Circumventing Concept Inhibition in Diffusion Models	擴散模型安全 (Honorable Mention)
Adversarial Diffusion Distillation	擴散模型蒸餾
Concept Sliders: LoRA Adaptors for Precise Control	可控生成
Grounding DINO: Marrying DINO with Grounded Pre-Training	開放集偵測
LGM: Large Multi-View Gaussian Model for 3D Content Creation	3D 生成
Sapiens: Foundation for Human Vision Models	人體基礎模型
SEA-RAFT: Simple, Efficient, Accurate RAFT for Optical Flow	光流估測
SiT: Exploring Flow and Diffusion-based Generative Models	生成模型
VideoMamba: State Space Model for Efficient Video Understanding	影片理解
DUSt3R: Geometric 3D Vision Made Easy	3D 重建