Model Agnostic Preference Optimization for Medical Image Segmentation
Not on the map. Why?
Abstract (arXiv)
Preference optimization offers a scalable supervision paradigm based on relative preference signals, yet prior attempts in medical image segmentation remain model-specific and rely on low-diversity prediction sampling. In this paper, we propose MAPO (Model-Agnostic Preference Optimization), a training framework that utilizes Dropout-driven stochastic segmentation hypotheses to construct preference-consistent gradients without direct ground-truth supervision. MAPO is fully architecture- and dimensionality-agnostic, supporting 2D/3D CNN and Transformer-based segmentation pipelines. Comprehensive evaluations across diverse medical datasets reveal that MAPO consistently enhances boundary adherence, reduces overfitting, and yields more stable optimization dynamics compared to conventional supervised training.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 42 of its 59 references that are in the map. First posted on arXiv 17 Dec 2025.
Builds on
42 of 59 references.
References · 42
- 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
- 2022Training language models to follow instructions with human feedbackcited by 15,971
- 2017Proximal Policy Optimization Algorithmscited by 15,215
- 2023Segment Anythingcited by 8,234
- 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
- 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,513
- 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
- 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
- 2022Constitutional AI: Harmlessness from AI Feedbackcited by 3,050
- 2016V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentationcited by 2,292
- 2021TransUNet: Transformers Make Strong Encoders for Medical Image Segmentationcited by 1,558
- 2018Attention U-Net: Learning Where to Look for the Pancreascited by 1,435
30 more
- 2021Swin-Unet: Unet-like Pure Transformer for Medical Image Segmentationcited by 1,222
- 2023A General Theoretical Paradigm to Understand Learning from Human Preferencescited by 912
- 2019Kvasir-SEG: A Segmented Polyp Datasetcited by 589
- 2024ORPO: Monolithic Preference Optimization without Reference Modelcited by 562
- 2023RRHF: Rank Responses to Align Language Models with Human Feedback without tearscited by 463
- 2019Uncertainty-aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentationcited by 379
- 2021Domain Adaptation for Medical Image Analysis: A Surveycited by 352
- 2018nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentationcited by 291
- 2023Segment Anything Model for Medical Image Analysis: an Experimental Studycited by 291
- 2017Tversky loss function for image segmentation using 3D fully convolutional deep networkscited by 228
- 2023Cross-Entropy Loss Functions: Theoretical Analysis and Applicationscited by 210
- 2020Can AI help in screening Viral and COVID-19 pneumonia?cited by 200
- 2018Boundary loss for highly unbalanced segmentationcited by 190
- 2019Models Genesis: Generic Autodidactic Models for 3D Medical Image Analysiscited by 159
- 2020Shape-aware Semi-supervised 3D Semantic Segmentation for Medical Imagescited by 152
- 2019Reducing the Hausdorff Distance in Medical Image Segmentation with Convolutional Neural Networkscited by 116
- 2021MISSFormer: An Effective Medical Image Segmentation Transformercited by 103
- 2023Large-Vocabulary Segmentation for Medical Images with Text Promptscited by 62
- 2021TransAttUnet: Multi-level Attention-guided U-Net with Transformer for Medical Image Segmentationcited by 60
- 2023SAM Fails to Segment Anything? -- SAM-Adapter: Adapting SAM in Underperformed Scenes: Camouflage, Shadow, Medical Image Segmentation, and Morecited by 53
- 2024Unleashing the Potential of SAM for Medical Adaptation via Hierarchical Decodingcited by 48
- 2022D-Former: A U-shaped Dilated Transformer for 3D Medical Image Segmentationcited by 35
- 2024Segment anything model 2: an application to 2D and 3D medical imagescited by 28
- 2023A Generalized Surface Loss for Reducing the Hausdorff Distance in Medical Imaging Segmentationcited by 13
- 2025Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentationcited by 12
- 2025Causal-SAM-LLM: Large Language Models as Causal Reasoners for Robust Medical Segmentationcited by 5
- 2024Annotation-Efficient Language Model Alignment via Diverse and Representative Response Textscited by 3
- 2025SAM-aware Test-time Adaptation for Universal Medical Image Segmentationcited by 3
- 2023AutoPET Challenge 2023: Sliding Window-based Optimization of U-Netcited by 2
- 2025SAMPO-Path: Segmentation Intent-Aligned Preference Optimization for Pathology Foundation Model Segmentationcited by 2
Built on by
Citing papers per year: 2026: 1.
1 paper.
Similar, not cited
Close in wording, no citation either way
- 2023Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective
- 2024SPA: Efficient User-Preference Alignment against Uncertainty in Medical Image Segmentation
- 2023Rethinking Data Perturbation and Model Stabilization for Semi-supervised Medical Image Segmentation
- 2026Camyla: Scaling Autonomous Research in Medical Image Segmentation
- 2025Generalist Models in Medical Image Segmentation: A Survey and Performance Comparison with Task-Specific Approaches
- 2023Ambiguous Medical Image Segmentation using Diffusion Models
- 2023Alternate Diverse Teaching for Semi-supervised Medical Image Segmentation
- 2026SegMoTE: Token-Level Mixture of Experts for Medical Image Segmentation