CVPR 2024 accepted papers on Label-efficient and robust learning
240 of the 2,716 papers accepted at CVPR 2024 (8.8%) are on Label-efficient and robust learning, compared with 14.2% at CVPR 2023. They fall into 9 topics.
See CVPR 2024 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Recognition and long-tail · 86 papers
- RoMa: Robust Dense Feature Matching (263 citations)
- Optimal Transport Aggregation for Visual Place Recognition (137 citations)
- Feature Re-Embedding: Towards Foundation Model-Level Performance in Computational Pathology (104 citations)
- Preserving Fairness Generalization in Deepfake Detection (94 citations)
- Is Ego Status All You Need for Open-Loop End-to-End Autonomous Driving? (78 citations)
- Implicit Discriminative Knowledge Learning for Visible-Infrared Person Re-Identification (76 citations)
- Morphological Prototyping for Unsupervised Slide Representation Learning in Computational Pathology (70 citations)
- Unleashing the Potential of SAM for Medical Adaptation via Hierarchical Decoding (70 citations)
- Shallow-Deep Collaborative Learning for Unsupervised Visible-Infrared Person Re-Identification (69 citations)
- BigGait: Learning Gait Representation You Want by Large Vision Models (58 citations)
- DePT: Decoupled Prompt Tuning (50 citations)
- MindBridge: A Cross-Subject Brain Decoding Framework (46 citations)
and 74 more in the interactive map.
Domain adaptation and generalization · 41 papers
- CFPL-FAS: Class Free Prompt Learning for Generalizable Face Anti-spoofing (79 citations)
- Source-Free Domain Adaptation with Frozen Multimodal Foundation Model (50 citations)
- CAT: Exploiting Inter-Class Dynamics for Domain Adaptive Object Detection (47 citations)
- Domain-Agnostic Mutual Prompting for Unsupervised Domain Adaptation (44 citations)
- Learning CNN on ViT: A Hybrid Model to Explicitly Class-specific Boundaries for Domain Adaptation (44 citations)
- Gradient Alignment for Cross-Domain Face Anti-Spoofing (41 citations)
- Single Domain Generalization for Crowd Counting (37 citations)
- Day-Night Cross-domain Vehicle Re-identification (36 citations)
- MAPSeg: Unified Unsupervised Domain Adaptation for Heterogeneous Medical Image Segmentation Based on 3D Masked Autoencoding and Pseudo-Labeling (35 citations)
- LEAD: Learning Decomposition for Source-free Universal Domain Adaptation (34 citations)
- Domain-Rectifying Adapter for Cross-Domain Few-Shot Segmentation (33 citations)
- Style Blind Domain Generalized Semantic Segmentation via Covariance Alignment and Semantic Consistence Contrastive Learning (33 citations)
and 29 more in the interactive map.
Zero-/few-shot and meta-learning · 22 papers
- Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data (1,178 citations)
- Telling Left from Right: Identifying Geometry-Aware Semantic Correspondence (48 citations)
- One-Prompt to Segment All Medical Images (46 citations)
- Visual-Augmented Dynamic Semantic Prototype for Generative Zero-Shot Learning (39 citations)
- Improving the Generalization of Segmentation Foundation Model under Distribution Shift via Weakly Supervised Adaptation (36 citations)
- APSeg: Auto-Prompt Network for Cross-Domain Few-Shot Semantic Segmentation (34 citations)
- Transductive Zero-Shot and Few-Shot CLIP (27 citations)
- Adapt Before Comparison: A New Perspective on Cross-Domain Few-Shot Segmentation (25 citations)
- Cross-Domain Few-Shot Segmentation via Iterative Support-Query Correspondence Mining (24 citations)
- Frozen Feature Augmentation for Few-Shot Image Classification (22 citations)
- Discriminative Sample-Guided and Parameter-Efficient Feature Space Adaptation for Cross-Domain Few-Shot Learning (18 citations)
- DeIL: Direct-and-Inverse CLIP for Open-World Few-Shot Learning (15 citations)
and 10 more in the interactive map.
Self-supervised and contrastive · 22 papers
- VoCo: A Simple-yet-Effective Volume Contrastive Learning Framework for 3D Medical Image Analysis (87 citations)
- Transfer CLIP for Generalizable Image Denoising (48 citations)
- Masked Autoencoders for Microscopy are Scalable Learners of Cellular Biology (46 citations)
- Transcriptomics-guided Slide Representation Learning in Computational Pathology (46 citations)
- SwitchLight: Co-design of Physics-driven Architecture and Pre-training Framework for Human Portrait Relighting (35 citations)
- ODCR: Orthogonal Decoupling Contrastive Regularization for Unpaired Image Dehazing (31 citations)
- Investigating and Mitigating the Side Effects of Noisy Views for Self-Supervised Clustering Algorithms in Practical Multi-View Scenarios (29 citations)
- Self-Supervised Facial Representation Learning with Facial Region Awareness (25 citations)
- LAFS: Landmark-based Facial Self-supervised Learning for Face Recognition (22 citations)
- Parameter Efficient Self-Supervised Geospatial Domain Adaptation (20 citations)
- Bootstrapping Autonomous Driving Radars with Self-Supervised Learning (17 citations)
- An Asymmetric Augmented Self-Supervised Learning Method for Unsupervised Fine-Grained Image Hashing (15 citations)
and 10 more in the interactive map.
OOD and anomaly detection · 18 papers
- RealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detection (198 citations)
- Transcending Forgery Specificity with Latent Space Augmentation for Generalizable Deepfake Detection (152 citations)
- PromptAD: Learning Prompts with only Normal Samples for Few-Shot Anomaly Detection (135 citations)
- Toward Generalist Anomaly Detection via In-context Residual Learning with Few-shot Sample Prompts (78 citations)
- Anomaly Heterogeneity Learning for Open-set Supervised Anomaly Detection (37 citations)
- Learning Transferable Negative Prompts for Out-of-Distribution Detection (32 citations)
- Hyperbolic Anomaly Detection (28 citations)
- ID-like Prompt Learning for Few-Shot Out-of-Distribution Detection (27 citations)
- Can Biases in ImageNet Models Explain Generalization? (21 citations)
- CORES: Convolutional Response-based Score for Out-of-distribution Detection (16 citations)
- Long-Tailed Anomaly Detection with Learnable Class Names (15 citations)
- Rethinking the Evaluation Protocol of Domain Generalization (13 citations)
and 6 more in the interactive map.
Test-time adaptation · 16 papers
- Efficient Test-Time Adaptation of Vision-Language Models (70 citations)
- Test-Time Domain Generalization for Face Anti-Spoofing (59 citations)
- Localization Is All You Evaluate: Data Leakage in Online Mapping Datasets and How to Fix It (22 citations)
- T4P: Test-Time Training of Trajectory Prediction via Masked Autoencoder and Actor-specific Token Memory (18 citations)
- GDA: Generalized Diffusion for Robust Test-time Adaptation (16 citations)
- Improved Self-Training for Test-Time Adaptation (15 citations)
- NC-TTT: A Noise Constrastive Approach for Test-Time Training (13 citations)
- TEA: Test-time Energy Adaptation (13 citations)
- Test-Time Adaptation for Depth Completion (11 citations)
- Any-Shift Prompting for Generalization over Distributions (8 citations)
- Backpropagation-free Network for 3D Test-time Adaptation (8 citations)
- Fast Adaptation for Human Pose Estimation via Meta-Optimization (8 citations)
and 4 more in the interactive map.
Knowledge distillation · 15 papers
- Logit Standardization in Knowledge Distillation (243 citations)
- CrossKD: Cross-Head Knowledge Distillation for Object Detection (132 citations)
- CLIP-KD: An Empirical Study of CLIP Model Distillation (65 citations)
- Scaled Decoupled Distillation (47 citations)
- FreeKD: Knowledge Distillation via Semantic Frequency Prompt (40 citations)
- On the Diversity and Realism of Distilled Dataset: An Efficient Dataset Distillation Paradigm (31 citations)
- KD-DETR: Knowledge Distillation for Detection Transformer with Consistent Distillation Points Sampling (26 citations)
- Small Scale Data-Free Knowledge Distillation (21 citations)
- NAYER: Noisy Layer Data Generation for Efficient and Effective Data-free Knowledge Distillation (18 citations)
- VkD: Improving Knowledge Distillation using Orthogonal Projections (17 citations)
- On the Road to Portability: Compressing End-to-End Motion Planner for Autonomous Driving (16 citations)
- Asymmetric Masked Distillation for Pre-Training Small Foundation Models (12 citations)
and 3 more in the interactive map.
Semi-supervised learning · 13 papers
- Adaptive Bidirectional Displacement for Semi-Supervised Medical Image Segmentation (110 citations)
- A Semi-supervised Nighttime Dehazing Baseline with Spatial-Frequency Aware and Realistic Brightness Constraint (74 citations)
- Constructing and Exploring Intermediate Domains in Mixed Domain Semi-supervised Medical Image Segmentation (45 citations)
- Unleashing Unlabeled Data: A Paradigm for Cross-View Geo-Localization (44 citations)
- Towards the Uncharted: Density-Descending Feature Perturbation for Semi-supervised Semantic Segmentation (41 citations)
- PH-Net: Semi-Supervised Breast Lesion Segmentation via Patch-wise Hardness (31 citations)
- CDMAD: Class-Distribution-Mismatch-Aware Debiasing for Class-Imbalanced Semi-Supervised Learning (20 citations)
- Systematic Comparison of Semi-supervised and Self-supervised Learning for Medical Image Classification (20 citations)
- Targeted Representation Alignment for Open-World Semi-Supervised Learning (18 citations)
- BEM: Balanced and Entropy-based Mix for Long-Tailed Semi-Supervised Learning (11 citations)
- Towards Generalizing to Unseen Domains with Few Labels (10 citations)
- Uncertainty-Guided Never-Ending Learning to Drive (6 citations)
and 1 more in the interactive map.
Noisy labels · 7 papers
- L2B: Learning to Bootstrap Robust Models for Combating Label Noise (20 citations)
- Learning with Structural Labels for Learning with Noisy Labels (20 citations)
- Learning Discriminative Dynamics with Label Corruption for Noisy Label Detection (16 citations)
- Estimating Noisy Class Posterior with Part-level Labels for Noisy Label Learning (15 citations)
- SURE: SUrvey REcipes for building reliable and robust deep networks (15 citations)
- CroSel: Cross Selection of Confident Pseudo Labels for Partial-Label Learning (11 citations)
- Unsupervised Deep Unrolling Networks for Phase Unwrapping (6 citations)