ICCV 2023 accepted papers on Label-efficient and robust learning
320 of the 2,156 papers accepted at ICCV 2023 (14.8%) are on Label-efficient and robust learning, compared with 16.9% at ICCV 2021. They fall into 9 topics.
See ICCV 2023 in the interactive map
Submitting to ICCV? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Recognition and long-tail · 88 papers
- VAD: Vectorized Scene Representation for Efficient Autonomous Driving (478 citations)
- EigenPlaces: Training Viewpoint Robust Models for Visual Place Recognition (167 citations)
- Visible-Infrared Person Re-Identification via Semantic Alignment and Affinity Inference (113 citations)
- Modality Unifying Network for Visible-Infrared Person Re-Identification (111 citations)
- Towards Grand Unified Representation Learning for Unsupervised Visible-Infrared Person Re-Identification (108 citations)
- DriveAdapter: Breaking the Coupling Barrier of Perception and Planning in End-to-End Autonomous Driving (107 citations)
- PivotNet: Vectorized Pivot Learning for End-to-end HD Map Construction (105 citations)
- Hidden Biases of End-to-End Driving Models (101 citations)
- Learning Gabor Texture Features for Fine-Grained Recognition (77 citations)
- Global Features are All You Need for Image Retrieval and Reranking (73 citations)
- AREA: Adaptive Reweighting via Effective Area for Long-Tailed Classification (71 citations)
- CDUL: CLIP-Driven Unsupervised Learning for Multi-Label Image Classification (71 citations)
and 76 more in the interactive map.
Domain adaptation and generalization · 75 papers
- Part-Aware Transformer for Generalizable Person Re-identification (153 citations)
- DomainDrop: Suppressing Domain-Sensitive Channels for Domain Generalization (72 citations)
- PADCLIP: Pseudo-labeling with Adaptive Debiasing in CLIP for Unsupervised Domain Adaptation (71 citations)
- Camera-Driven Representation Learning for Unsupervised Domain Adaptive Person Re-identification (59 citations)
- CMDA: Cross-Modality Domain Adaptation for Nighttime Semantic Segmentation (55 citations)
- Towards Unsupervised Domain Generalization for Face Anti-Spoofing (55 citations)
- One-Shot Generative Domain Adaptation (53 citations)
- Diffusion-based Image Translation with Label Guidance for Domain Adaptive Semantic Segmentation (51 citations)
- GraphEcho: Graph-Driven Unsupervised Domain Adaptation for Echocardiogram Video Segmentation (48 citations)
- Texture Learning Domain Randomization for Domain Generalized Segmentation (48 citations)
- Controllable Guide-Space for Generalizable Face Forgery Detection (41 citations)
- Generalizable Decision Boundaries: Dualistic Meta-Learning for Open Set Domain Generalization (40 citations)
and 63 more in the interactive map.
Self-supervised and contrastive · 34 papers
- Scale-MAE: A Scale-Aware Masked Autoencoder for Multiscale Geospatial Representation Learning (357 citations)
- Forecast-MAE: Self-supervised Pre-training for Motion Forecasting with Masked Autoencoders (125 citations)
- An Embarrassingly Simple Backdoor Attack on Self-supervised Learning (69 citations)
- Pre-Training-Free Image Manipulation Localization through Non-Mutually Exclusive Contrastive Learning (62 citations)
- Audiovisual Masked Autoencoders (54 citations)
- Self-Supervised Character-to-Character Distillation for Text Recognition (50 citations)
- Event Camera Data Pre-training (44 citations)
- Self-supervised Cross-view Representation Reconstruction for Change Captioning (43 citations)
- EfficientTrain: Exploring Generalized Curriculum Learning for Training Visual Backbones (39 citations)
- Geometrized Transformer for Self-Supervised Homography Estimation (37 citations)
- Identity-Seeking Self-Supervised Representation Learning for Generalizable Person Re-Identification (35 citations)
- Anatomical Invariance Modeling and Semantic Alignment for Self-supervised Learning in 3D Medical Image Analysis (33 citations)
and 22 more in the interactive map.
Semi-supervised learning · 32 papers
- XNet: Wavelet-Based Low and High Frequency Fusion Networks for Fully- and Semi-Supervised Semantic Segmentation of Biomedical Images (155 citations)
- Parametric Classification for Generalized Category Discovery: A Baseline Study (144 citations)
- CauSSL: Causality-inspired Semi-supervised Learning for Medical Image Segmentation (125 citations)
- Learning Semi-supervised Gaussian Mixture Models for Generalized Category Discovery (100 citations)
- Dual Pseudo-Labels Interactive Self-Training for Semi-Supervised Visible-Infrared Person Re-Identification (76 citations)
- Enhanced Soft Label for Semi-Supervised Semantic Segmentation (76 citations)
- Logic-induced Diagnostic Reasoning for Semi-supervised Semantic Segmentation (68 citations)
- Diverse Cotraining Makes Strong Semi-Supervised Segmentor (60 citations)
- Pseudo-label Alignment for Semi-supervised Instance Segmentation (43 citations)
- Harvard Glaucoma Detection and Progression: A Multimodal Multitask Dataset and Generalization-Reinforced Semi-Supervised Learning (38 citations)
- IOMatch: Simplifying Open-Set Semi-Supervised Learning with Joint Inliers and Outliers Utilization (38 citations)
- Calibrating Uncertainty for Semi-Supervised Crowd Counting (31 citations)
and 20 more in the interactive map.
Knowledge distillation · 28 papers
- From Knowledge Distillation to Self-Knowledge Distillation: A Unified Approach with Normalized Loss and Customized Soft Labels (176 citations)
- Multiple Instance Learning Framework with Masked Hard Instance Mining for Whole Slide Image Classification (138 citations)
- DREAM: Efficient Dataset Distillation by Representative Matching (93 citations)
- Bridging Cross-task Protocol Inconsistency for Distillation in Dense Object Detection (89 citations)
- Remembering Normality: Memory-guided Knowledge Distillation for Unsupervised Anomaly Detection (83 citations)
- DataDAM: Efficient Dataset Distillation with Attention Matching (63 citations)
- Automated Knowledge Distillation via Monte Carlo Tree Search (52 citations)
- DETRDistill: A Universal Knowledge Distillation Framework for DETR-families (52 citations)
- A Good Student is Cooperative and Reliable: CNN-Transformer Collaborative Learning for Semantic Segmentation (41 citations)
- Dual Learning with Dynamic Knowledge Distillation for Partially Relevant Video Retrieval (36 citations)
- Class-relation Knowledge Distillation for Novel Class Discovery (31 citations)
- Multi-Label Knowledge Distillation (29 citations)
and 16 more in the interactive map.
OOD and anomaly detection · 25 papers
- Unsupervised Surface Anomaly Detection with Diffusion Probabilistic Model (153 citations)
- PNI : Industrial Anomaly Detection using Position and Neighborhood Information (141 citations)
- FastRecon: Few-shot Industrial Anomaly Detection via Fast Feature Reconstruction (128 citations)
- CLIPN for Zero-Shot OOD Detection: Teaching CLIP to Say No (125 citations)
- Normalizing Flows for Human Pose Anomaly Detection (112 citations)
- Anomaly Detection Under Distribution Shift (73 citations)
- Nearest Neighbor Guidance for Out-of-Distribution Detection (62 citations)
- Residual Pattern Learning for Pixel-Wise Out-of-Distribution Detection in Semantic Segmentation (60 citations)
- Template-guided Hierarchical Feature Restoration for Anomaly Detection (59 citations)
- Focus the Discrepancy: Intra- and Inter-Correlation Learning for Image Anomaly Detection (42 citations)
- Inter-Realization Channels: Unsupervised Anomaly Detection Beyond One-Class Classification (39 citations)
- Anomaly Detection using Score-based Perturbation Resilience (27 citations)
and 13 more in the interactive map.
Zero-/few-shot and meta-learning · 17 papers
- Metric3D: Towards Zero-shot Metric 3D Prediction from A Single Image (314 citations)
- Class-Aware Patch Embedding Adaptation for Few-Shot Image Classification (71 citations)
- Frequency Guidance Matters in Few-Shot Learning (67 citations)
- PEANUT: Predicting and Navigating to Unseen Targets (40 citations)
- Similarity Min-Max: Zero-Shot Day-Night Domain Adaptation (39 citations)
- Hyperbolic Audio-visual Zero-shot Learning (28 citations)
- Prototypes-oriented Transductive Few-shot Learning with Conditional Transport (26 citations)
- Distilled Reverse Attention Network for Open-world Compositional Zero-Shot Learning (23 citations)
- Image-Free Classifier Injection for Zero-Shot Classification (23 citations)
- Generalized Sum Pooling for Metric Learning (7 citations)
- Continual Zero-Shot Learning through Semantically Guided Generative Random Walks (5 citations)
- Benchmarking Low-Shot Robustness to Natural Distribution Shifts (4 citations)
and 5 more in the interactive map.
Test-time adaptation · 11 papers
- DomainAdaptor: A Novel Approach to Test-time Adaptation (56 citations)
- DETA: Denoised Task Adaptation for Few-Shot Learning (42 citations)
- On the Robustness of Open-World Test-Time Training: Self-Training with Dynamic Prototype Expansion (27 citations)
- Towards Open-Set Test-Time Adaptation Utilizing the Wisdom of Crowds in Entropy Minimization (23 citations)
- CAFA: Class-Aware Feature Alignment for Test-Time Adaptation (20 citations)
- Test Time Adaptation for Blind Image Quality Assessment (20 citations)
- Label Shift Adapter for Test-Time Adaptation under Covariate and Label Shifts (19 citations)
- Point-TTA: Test-Time Adaptation for Point Cloud Registration Using Multitask Meta-Auxiliary Learning (19 citations)
- ClusT3: Information Invariant Test-Time Training (18 citations)
- Test-time Personalizable Forecasting of 3D Human Poses (7 citations)
- Rapid Network Adaptation: Learning to Adapt Neural Networks Using Test-Time Feedback
Noisy labels · 10 papers
- Combating Noisy Labels with Sample Selection by Mining High-Discrepancy Examples (72 citations)
- LA-Net: Landmark-Aware Learning for Reliable Facial Expression Recognition under Label Noise (56 citations)
- Label-Noise Learning with Intrinsically Long-Tailed Data (31 citations)
- When Noisy Labels Meet Long Tail Dilemmas: A Representation Calibration Method (29 citations)
- RankMatch: Fostering Confidence and Consistency in Learning with Noisy Labels (25 citations)
- PADDLES: Phase-Amplitude Spectrum Disentangled Early Stopping for Learning with Noisy Labels (21 citations)
- SILT: Shadow-Aware Iterative Label Tuning for Learning to Detect Shadows from Noisy Labels (20 citations)
- Late Stopping: Avoiding Confidently Learning from Mislabeled Examples (13 citations)
- Sample-wise Label Confidence Incorporation for Learning with Noisy Labels (9 citations)
- Prototypical Mixing and Retrieval-Based Refinement for Label Noise-Resistant Image Retrieval (8 citations)
