CVPR2025 medical ai papers

Partially from https://github.com/MedAIerHHL/CVPR-MIA

Additions or corrections are welcome in Issues!

format: Title + Paper Link + Code Link. Your contributions are always welcome!

World Model

• EchoWorld: Learning Motion-Aware World Models for Echocardiography Probe Guidance [Paper][Code]

• CheXWorld: Exploring Image World Modeling for Radiograph Representation Learning [Paper][Code]

Image Segmentation (图像分割)

• Interactive Medical Image Segmentation: A Benchmark Dataset and Baseline [Paper] [Code]

• Steady Progress Beats Stagnation: Mutual Aid of Foundation and Conventional Models in Mixed Domain Semi-Supervised Medical Image Segmentation [Paper] [Code]

• DyCON: Dynamic Uncertainty-aware Consistency and Contrastive Learning for Semi-supervised Medical Image Segmentation [Paper]
• EffiDec3D: An Optimized Decoder for High-Performance and Efficient 3D Medical Image Segmentation
• Test-Time Domain Generalization via Universe Learning: A Multi-Graph Matching Approach for Medical Image Segmentation [Paper] [Code]
• nnWNet: Rethinking the Use of Transformers in Biomedical Image Segmentation and Calling for a Unified Evaluation Benchmark. [Paper][Code]
• Interactive Medical Image Segmentation: A Benchmark Dataset and Baseline. [Paper][Code]
• Advancing Generalizable Tumor Segmentation with Anomaly.Aware Open-Vocabulary Attention Maps and Frozen FoundationDiffusion Models.
• LesionLocator: Zero-Shot Universal Tumor Segmentation and Tracking in 3D Whole-Body Imaging. [Paper][Code]
• Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentation. [Paper][Code]
• Boost the Inference with Co-training: A Depth-guided Mutual Learning Framework for Semi-supervised Medical Polyp Segmentation (RD-Net) [Code]

Vision-Language Model (视觉-语言)

• VILA-M3: Enhancing Vision-Language Models with Medical Expert Knowledge [paper] [code] (based on [MONAI]

• BIOMEDICA: An Open Biomedical Image-Caption Archive with Vision-Language Models derived from Scientific Literature [paper] [project]

• BiomedCoOp: Learning to Prompt for Biomedical Vision-Language Models [paper] [code]

• MIMO: A medical vision language model with visual referring multimodal input and pixel grounding multimodal output [Paper] [code]
• Towards a Multimodal Large Language Model with Pixel-Level Insight for Biomedicine [Paper] [code] (AAAI-2025)

• Bringing CLIP to the Clinic: Dynamic Soft Labels and Negation-Aware Learning for Medical Analysis
• Alignment, Mining and Fusion: Representation Alignment with Hard Negative Mining and Selective Knowledge Fusion for Medical Visual Question Answering
• Enhanced Contrastive Learning with Multi-view Longitudinal Data for Chest X-ray Report Generation. [Paper][Code]

• MedUnifier: Unifying Vision-and-Language Pre-training on Medical Data with Vision Generation Task using Discrete Visual Representations. [Paper][Code]

Multi modality (多模态)

• Multi-modal Medical Diagnosis via Large-small Model Collaboration [paper] [code]

• MM-OR: A Large Multimodal Operating Room Dataset for Semantic Understanding of High-Intensity Surgical Environments [paper] [code]

• Multi-modal Vision Pre-training for Medical Image Analysis [paper] [code]

• STiL: Semi-supervised Tabular-Image Learning for Comprehensive Task-Relevant Information Exploration in Multimodal Classification [paper] [code]

• Knowledge Bridger: Towards Training-Free Missing Modality Completion [paper] [code]

• Prompt2Perturb (P2P): Text-Guided Diffusion-Based Adversarial Attacks on Breast Ultrasound Images [paper] [code]

Computational Pathology (计算病理)

• Fast and Accurate Gigapixel Pathological Image Classification with Hierarchical Distillation Multi-Instance LearningComputational Pathology. [Paper][Code]
• FOCUS: Knowledge-enhanced Adaptive Visual Compression for Few-shot Whole Slide Image Classification. [Paper][Code][推送]

• Distilled Prompt Learning for Incomplete Multimodal Survival Prediction. [Paper] [Code]
• Fast and Accurate Gigapixel Pathological Image Classification with Hierarchical Distillation Multi-Instance Learning. [Paper][Code]
• SlideChat: A Large Vision-Language Assistant for Whole-Slide Pathology Image Understanding. [Paper][Code]
• 2DMamba: Efficient State Space Model for Image Representation with Applications on Giga-Pixel Whole Slide Image Classification. [Paper][Code]
• CPath-Omni: A Unified Multimodal Foundation Model for Patch and Whole Slide Image Analysis in Computational Pathology. [Paper][Code]
• MERGE: Multi-faceted Hierarchical Graph-based GNN for Gene Expression Prediction from Whole Slide Histopathology Images. [Paper][Code]
• HistoFS: Non-IID Histopathologic Whole Slide Image Classification via Federated Style Transfer with RoI-Preserving. [Paper][Code]
• M3amba: Memory Mamba is All You Need for Whole Slide Image Classification. [Paper][Code]
• Advancing Multiple Instance Learning with Continual Learning for Whole Slide Imaging. [Paper][Code]
• BioX-CPath: Biologically-driven Explainable Diagnostics for Multistain IHC Computational Pathology. [Paper][Code]
• Multi-Resolution Pathology-Language Pre-training Model with Text-Guided Visual Representation. [Paper][Code]
• TopoCellGen: Generating Histopathology Cell Topology with a Diffusion Model. [Paper][Code]
• Multi-modal Topology-embedded Graph Learning for Spatially Resolved Genes Prediction from Pathology Images with Prior Gene Similarity Information. [Paper][Code]
• Robust Multimodal Survival Prediction with the Latent Differentiation Conditional Variational AutoEncoder. [Paper][Code]
• MExD: An Expert-Infused Diffusion Model for Whole-Slide Image Classification. [Paper][Code]
• Learning Heterogeneous Tissues with Mixture of Experts for Gigapixel Whole Slide Images. [Paper][Code]
• Unsupervised Foundation Model-Agnostic Slide-Level Representation Learning. [Paper][Code]
• WISE: A Framework for Gigapixel Whole-Slide-Image Lossless Compression. [Paper][Code]

Others

• Blood Flow Speed Estimation with Optical Coherence Tomography Angiography Images [Paper] [Code]

• MuTri: Multi-view Tri-alignment for OCT to OCTA 3D Image Translation [Paper][Code]

• dFLMoE: Decentralized Federated Learning via Mixture of Experts for Medical Data Analysis [Paper]
• Q-PART: Quasi-Periodic Adaptive Regression with Test-time Training for Pediatric Left Ventricular Ejection Fraction Regression.
• Towards All-in-One Medical Image Re-Identification [Paper]