CVPR 2025 accepted papers on Label-efficient and robust learning
222 of the 2,871 papers accepted at CVPR 2025 (7.7%) are on Label-efficient and robust learning, compared with 8.8% at CVPR 2024. They fall into 9 topics.
See CVPR 2025 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 · 70 papers
- From Poses to Identity: Training-Free Person Re-Identification via Feature Centralization (23 citations)
- FreqDebias: Towards Generalizable Deepfake Detection via Consistency-Driven Frequency Debiasing (18 citations)
- SeCap: Self-Calibrating and Adaptive Prompts for Cross-view Person Re-Identification in Aerial-Ground Networks (16 citations)
- CLOC: Contrastive Learning for Ordinal Classification with Multi-Margin N-pair Loss (15 citations)
- Finer-CAM: Spotting the Difference Reveals Finer Details for Visual Explanation (14 citations)
- A Tale of Two Classes: Adapting Supervised Contrastive Learning to Binary Imbalanced Datasets (12 citations)
- Cheb-GR: Rethinking K-nearest Neighbor Search in Re-ranking for Person Re-identification (12 citations)
- DIFFER: Disentangling Identity Features via Semantic Cues for Clothes-Changing Person Re-ID (12 citations)
- Enhancing Testing-Time Robustness for Trusted Multi-View Classification in the Wild (12 citations)
- Feature Spectrum Learning for Remote Sensing Change Detection (12 citations)
- Bridging Past and Future: End-to-End Autonomous Driving with Historical Prediction and Planning (11 citations)
- Learning Heterogeneous Tissues with Mixture of Experts for Gigapixel Whole Slide Images (11 citations)
and 58 more in the interactive map.
Domain adaptation and generalization · 36 papers
- FisherTune: Fisher-Guided Robust Tuning of Vision Foundation Models for Domain Generalized Segmentation (16 citations)
- Domain Generalization in CLIP via Learning with Diverse Text Prompts (12 citations)
- Large Self-Supervised Models Bridge the Gap in Domain Adaptive Object Detection (10 citations)
- NightAdapter: Learning a Frequency Adapter for Generalizable Night-time Scene Segmentation (9 citations)
- SoMA: Singular Value Decomposed Minor Components Adaptation for Domain Generalizable Representation Learning (9 citations)
- Universal Domain Adaptation for Semantic Segmentation (9 citations)
- Dual-Agent Optimization framework for Cross-Domain Few-Shot Segmentation (8 citations)
- EventFly: Event Camera Perception from Ground to the Sky (7 citations)
- Multi-Granularity Class Prototype Topology Distillation for Class-Incremental Source-Free Unsupervised Domain Adaptation (7 citations)
- Optimal Transport-Guided Source-Free Adaptation for Face Anti-Spoofing (7 citations)
- Preserving Clusters in Prompt Learning for Unsupervised Domain Adaptation (7 citations)
- SemiDAViL: Semi-supervised Domain Adaptation with Vision-Language Guidance for Semantic Segmentation (7 citations)
and 24 more in the interactive map.
Semi-supervised learning · 26 papers
- DyCON: Dynamic Uncertainty-aware Consistency and Contrastive Learning for Semi-supervised Medical Image Segmentation (36 citations)
- beta-FFT: Nonlinear Interpolation and Differentiated Training Strategies for Semi-Supervised Medical Image Segmentation (16 citations)
- Annotation Ambiguity Aware Semi-Supervised Medical Image Segmentation (13 citations)
- Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentation (13 citations)
- A Semantic Knowledge Complementarity based Decoupling Framework for Semi-supervised Class-imbalanced Medical Image Segmentation (11 citations)
- CGMatch: A Different Perspective of Semi-supervised Learning (10 citations)
- Steady Progress Beats Stagnation: Mutual Aid of Foundation and Conventional Models in Mixed Domain Semi-Supervised Medical Image Segmentation (10 citations)
- Point-to-Region Loss for Semi-Supervised Point-Based Crowd Counting (8 citations)
- Taste More, Taste Better: Diverse Data and Strong Model Boost Semi-Supervised Crowd Counting (8 citations)
- Boost the Inference with Co-training: A Depth-guided Mutual Learning Framework for Semi-supervised Medical Polyp Segmentation (5 citations)
- Detect Any Mirrors: Boosting Learning Reliability on Large-Scale Unlabeled Data with an Iterative Data Engine (5 citations)
- Improving Semi-Supervised Semantic Segmentation with Sliced-Wasserstein Feature Alignment and Uniformity (5 citations)
and 14 more in the interactive map.
OOD and anomaly detection · 26 papers
- Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection (103 citations)
- Exploring Intrinsic Normal Prototypes within a Single Image for Universal Anomaly Detection (50 citations)
- UniVAD: A Training-free Unified Model for Few-shot Visual Anomaly Detection (30 citations)
- UniNet: A Contrastive Learning-guided Unified Framework with Feature Selection for Anomaly Detection (29 citations)
- D^3: Scaling Up Deepfake Detection by Learning from Discrepancy (13 citations)
- Detecting Out-of-Distribution Through the Lens of Neural Collapse (11 citations)
- OpenMIBOOD: Open Medical Imaging Benchmarks for Out-Of-Distribution Detection (10 citations)
- DPU: Dynamic Prototype Updating for Multimodal Out-of-Distribution Detection (9 citations)
- Leveraging Perturbation Robustness to Enhance Out-of-Distribution Detection (9 citations)
- On the Generalization of Handwritten Text Recognition Models (9 citations)
- OODD: Test-time Out-of-Distribution Detection with Dynamic Dictionary (9 citations)
- Overcoming Shortcut Problem in VLM for Robust Out-of-Distribution Detection (8 citations)
and 14 more in the interactive map.
Self-supervised and contrastive · 21 papers
- AnySat: One Earth Observation Model for Many Resolutions, Scales, and Modalities (28 citations)
- Revisiting MAE Pre-training for 3D Medical Image Segmentation (19 citations)
- Unsupervised Foundation Model-Agnostic Slide-Level Representation Learning (12 citations)
- FSFM: A Generalizable Face Security Foundation Model via Self-Supervised Facial Representation Learning (11 citations)
- Generalized Recorrupted-to-Recorrupted: Self-Supervised Learning Beyond Gaussian Noise (7 citations)
- AeSPa : Attention-guided Self-supervised Parallel Imaging for MRI Reconstruction (6 citations)
- VoteFlow: Enforcing Local Rigidity in Self-Supervised Scene Flow (6 citations)
- CheXWorld: Exploring Image World Modeling for Radiograph Representation Learning (4 citations)
- Synthetic-to-Real Self-supervised Robust Depth Estimation via Learning with Motion and Structure Priors (4 citations)
- A Regularization-Guided Equivariant Approach for Image Restoration (3 citations)
- UCM-VeID V2: A Richer Dataset and A Pre-training Method for UAV Cross-Modality Vehicle Re-Identification (3 citations)
- From Prototypes to General Distributions: An Efficient Curriculum for Masked Image Modeling (2 citations)
and 9 more in the interactive map.
Zero-/few-shot and meta-learning · 18 papers
- FoundationStereo: Zero-Shot Stereo Matching (98 citations)
- Stereo Anywhere: Robust Zero-Shot Deep Stereo Matching Even Where Either Stereo or Mono Fail (26 citations)
- vesselFM: A Foundation Model for Universal 3D Blood Vessel Segmentation (23 citations)
- Logits DeConfusion with CLIP for Few-Shot Learning (18 citations)
- On the Out-Of-Distribution Generalization of Large Multimodal Models (7 citations)
- Few-shot Personalized Scanpath Prediction (6 citations)
- Few-Shot Recognition via Stage-Wise Retrieval-Augmented Finetuning (5 citations)
- Forensic Self-Descriptions Are All You Need for Zero-Shot Detection, Open-Set Source Attribution, and Clustering of AI-generated Images (5 citations)
- ZeroVO: Visual Odometry with Minimal Assumptions (5 citations)
- Navigating the Unseen: Zero-shot Scene Graph Generation via Capsule-Based Equivariant Features (4 citations)
- OSLoPrompt: Bridging Low-Supervision Challenges and Open-Set Domain Generalization in CLIP (4 citations)
- Cross-View Completion Models are Zero-shot Correspondence Estimators (3 citations)
and 6 more in the interactive map.
Test-time adaptation · 12 papers
- Test-Time Domain Generalization via Universe Learning: A Multi-Graph Matching Approach for Medical Image Segmentation (7 citations)
- MetaWriter: Personalized Handwritten Text Recognition Using Meta-Learned Prompt Tuning (6 citations)
- Everything to the Synthetic: Diffusion-driven Test-time Adaptation via Synthetic-Domain Alignment (5 citations)
- Free on the Fly: Enhancing Flexibility in Test-Time Adaptation with Online EM (4 citations)
- Sufficient Invariant Learning for Distribution Shift (3 citations)
- Effortless Active Labeling for Long-Term Test-Time Adaptation (2 citations)
- Hierarchical Knowledge Prompt Tuning for Multi-task Test-Time Adaptation (2 citations)
- Tartan IMU: A Light Foundation Model for Inertial Positioning in Robotics (2 citations)
- Test-time Augmentation Improves Efficiency in Conformal Prediction (2 citations)
- PURA: Parameter Update-Recovery Test-Time Adaption for RGB-T Tracking (1 citation)
- TIDE: Training Locally Interpretable Domain Generalization Models Enables Test-time Correction (1 citation)
- AdMiT: Adaptive Multi-Source Tuning in Dynamic Environments
Noisy labels · 8 papers
- ROLL: Robust Noisy Pseudo-label Learning for Multi-View Clustering with Noisy Correspondence (14 citations)
- Learning with Noisy Triplet Correspondence for Composed Image Retrieval (9 citations)
- Theory-Inspired Deep Multi-View Multi-Label Learning with Incomplete Views and Noisy Labels (6 citations)
- D2SP: Dynamic Dual-Stage Purification Framework for Dual Noise Mitigation in Vision-based Affective Recognition. (5 citations)
- Directional Label Diffusion Model for Learning from Noisy Labels (4 citations)
- DiN: Diffusion Model for Robust Medical VQA with Semantic Noisy Labels (2 citations)
- Minding Fuzzy Regions: A Data-driven Alternating Learning Paradigm for Stable Lesion Segmentation (2 citations)
- The Impact Label Noise and Choice of Threshold has on Cross-Entropy and Soft-Dice in Image Segmentation (2 citations)
Knowledge distillation · 5 papers
- Accurate Scene Text Recognition with Efficient Model Scaling and Cloze Self-Distillation (5 citations)
- CustomKD: Customizing Large Vision Foundation for Edge Model Improvement via Knowledge Distillation (5 citations)
- Curriculum Coarse-to-Fine Selection for High-IPC Dataset Distillation (3 citations)
- Emphasizing Discriminative Features for Dataset Distillation in Complex Scenarios (2 citations)
- What Makes a Good Dataset for Knowledge Distillation?