HUR-MACL: High-Uncertainty Region-Guided Multi-Architecture Collaborative Learning for Head and Neck Multi-Organ Segmentation

2026-01-08

arXiv.org

AI Summary

This work addresses the challenge of accurately segmenting small and morphologically complex organs-at-risk in head and neck radiotherapy, where existing hybrid models suffer from functional redundancy and weak collaborative mechanisms. To overcome this limitation, we propose a high-uncertainty-region-guided multi-architecture collaborative learning framework. The method dynamically identifies regions with high segmentation uncertainty and leverages them to guide synergistic optimization between Vision Mamba and deformable convolutional networks in critical areas. Furthermore, a heterogeneous feature distillation loss is introduced to facilitate complementary learning across distinct architectures. By effectively integrating the strengths of each component while avoiding functional overlap, the proposed approach achieves state-of-the-art performance on two public benchmarks and one private dataset.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Accurate segmentation of organs at risk in the head and neck is essential for radiation therapy, yet deep learning models often fail on small, complexly shaped organs. While hybrid architectures that combine different models show promise, they typically just concatenate features without exploiting the unique strengths of each component. This results in functional overlap and limited segmentation accuracy. To address these issues, we propose a high uncertainty region-guided multi-architecture collaborative learning (HUR-MACL) model for multi-organ segmentation in the head and neck. This model adaptively identifies high uncertainty regions using a convolutional neural network, and for these regions, Vision Mamba as well as Deformable CNN are utilized to jointly improve their segmentation accuracy. Additionally, a heterogeneous feature distillation loss was proposed to promote collaborative learning between the two architectures in high uncertainty regions to further enhance performance. Our method achieves SOTA results on two public datasets and one private dataset.

Problem

• head and neck organ segmentation
• small organs
• complex-shaped organs
• multi-architecture fusion
• segmentation accuracy

Innovation

• High-Uncertainty Region
• Multi-Architecture Collaborative Learning
• Vision Mamba
• Deformable CNN
• Heterogeneous Feature Distillation