Model Agnostic Preference Optimization for Medical Image Segmentation

TL;DR

This paper introduces MAPO, a model-agnostic framework for medical image segmentation that leverages preference signals and stochastic hypotheses to improve boundary accuracy and training stability without relying on ground-truth labels.

Contribution

MAPO is a novel, architecture-agnostic training method that uses Dropout-driven hypotheses for preference-based supervision in medical image segmentation.

Findings

Enhances boundary adherence in segmentation results

Reduces overfitting during training

Provides more stable optimization dynamics

Abstract

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.