Mutual Distillation of Dual-Foundation Models for Semi-Supervised PET/CT Segmentation
Fuyou Mao, Beining Wu, Yanfeng Jiang, Bohan Xu, Lixin Lin, Naye Ji, Hao Zhang, Yan Tang
Why It Matters
What makes this one worth your time
This work addresses the challenge of high annotation costs in medical imaging, providing an efficient method for improving segmentation performance in oncology applications.
MuDuo leverages mutual distillation to enhance PET/CT segmentation with limited labeled data.
Summary
The paper presents a mutual distillation framework called MuDuo that integrates structural and functional foundation models for semi-supervised segmentation of PET/CT images, achieving state-of-the-art results with minimal labeled data.
Key contributions
- Introduction of the MuDuo mutual distillation framework for PET/CT segmentation.
- Demonstration of state-of-the-art performance on the AutoPET dataset with only 5 labeled cases.
- Utilization of SAM-Med3D and SegAnyPET models for knowledge distillation.
Notable insights
- The framework effectively combines knowledge from modality-specific models, which may improve generalization in semi-supervised settings.
- Eliminating the need for manual prompts could streamline the segmentation process in clinical workflows.
Possible limitations
- Not stated in the abstract.
Abstract
arXiv:2606.15611v1 Announce Type: cross Abstract: Organ segmentation from PET/CT is critical for quantitative analysis and radiotherapy planning in oncology. To ease the high annotation cost of PET/CT segmentation, semi-supervised learning (SSL) provides a practical and effective solution for developing deep models with limited labeled data. Recent developments in visual foundation models have demonstrated remarkable adaptability with improved efficiency. In this work, we propose a mutual distillation framework that seamlessly exploits both structural and functional foundation models, which act as modality-specific generalists for distilling knowledge from structural CT and metabolic PET imaging. By bridging the gap between the task-specific precision of student models and the segmentation priors of generalist foundation models, we propose \textbf{MuDuo}, a mutual distillation framework that synergistically leverages SAM-Med3D for CT and SegAnyPET for PET to distill their knowledge into a lightweight student network. Our approach eliminates the need for manual prompts while maximizing the utility of unlabeled data for automatic segmentation, achieving state-of-the-art performance on the AutoPET dataset with only 5 labeled cases. Our source code is available at https://github.com/Wu-beining/MuDuo.