BAP-MOS: Bandit-Based Adaptive Prompting for Boundary-Sensitive Multi-Organ Segmentation

2026-08-08

AI Summary

This work addresses the challenge of imprecise boundary segmentation in multi-organ ultrasound images, where anatomical adjacency leads to significant local boundary errors despite high Dice scores. To tackle this, the authors propose the BAP-MOS framework, which formulates prompt selection as an organ-specific multi-armed bandit problem. By employing a dual-loop mechanism—outer-layer Tree-structured Parzen Estimator (TPE) and inner-layer UCB-Tuned—the method adaptively optimizes only the mask decoder while keeping the image and prompt encoders frozen, thereby achieving boundary-sensitive prompting. Key innovations include the multi-armed bandit–based prompt selection strategy and an organ-scale negative prompt loop. Without altering the base model backbone, BAP-MOS achieves a Dice score of 0.982, HD95 of 0.482, and MSD of 0.204 on the TRUS dataset, reducing HD95 and MSD by 48% and 45%, respectively, over the strongest baseline, and demonstrates strong generalization on the PFUS1 dataset.

Abstract

Multi-organ ultrasound segmentation remains challenging when anatomically adjacent structures must be delineated jointly, as localized boundary errors can persist even when Dice scores are high. To address these challenges, we propose Boundary-Adaptive Prompting for Multi-Organ Segmentation (BAP-MOS), a closed-loop adaptive prompting framework. BAP-MOS formulates prompt selection as an organ-specific multi-armed bandit problem over box, point, and combined prompts. An outer Tree-structured Parzen Estimator (TPE) loop selects the prompt-selection parameter vector, while an inner UCB-Tuned loop adapts per-organ prompt preferences during fine-tuning using a bounded Dice--MSD--HD95 validation-probe reward. The framework further introduces an organ-scaled negative prompt ring to adapt sparse prompt geometry across anatomical scales, while keeping the image and prompt encoders frozen and updating only the mask decoder. We evaluate BAP-MOS on pooled prostate-region TRUS cohorts against U-Net, nnU-Net, MedSAM, fixed-prompt SAM/MedSAM, and adaptive policy variants. On this benchmark, BAP-MOS achieves Dice 0.982, HD95 0.482, and MSD 0.204, reducing HD95 by approximately 48% and MSD by 45% relative to the strongest conventional baseline. To verify the generalization ability of the framework, we tested it on the external PFUS1 pelvic-floor ultrasound corpus using MedSAM and its adaptive strategy variants, and the results were good. These results support adaptive prompt allocation as an effective mechanism for improving boundary-sensitive multi-organ ultrasound segmentation without modifying the foundation-model backbone. Source Code is available at: https://github.com/SatvikPraveen/BAP-MOS

Problem

• multi-organ segmentation
• boundary-sensitive
• ultrasound segmentation
• anatomically adjacent structures
• localized boundary errors

Innovation

• adaptive prompting
• multi-armed bandit
• boundary-sensitive segmentation
• ultrasound multi-organ segmentation
• foundation model adaptation