A Unified Framework for Foreground and Anonymization Area Segmentation in CT and MRI Data

2025-01-08

AI Summary

To address critical challenges in self-supervised learning (SSL) for 3D medical imaging (CT/MRI)—including inaccurate foreground segmentation, supervision contamination from anonymized regions, and the trade-off between cross-center privacy preservation and computational efficiency—this work proposes the first unified dual-branch 3D segmentation framework that jointly models foreground segmentation and anonymized region identification. The method employs a deep learning architecture specifically designed for volumetric data, integrating robust training strategies and cross-modal generalization mechanisms to avoid supervision corruption caused by anonymized regions—common in reconstruction-based SSL. Evaluated on multi-center CT and MRI datasets, it achieves >99.5% Dice for foreground segmentation and >98.5% Dice for anonymized region segmentation, demonstrating strong stability and generalizability. The code and pre-trained weights are publicly released, enabling privacy-safe, computationally efficient, and fully reproducible medical SSL pretraining.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Unsupervised & Self-Supervised Learning
• Natural Language Processing: Safety and Robustness

Application Category

• Security and Privacy: Data transparency and provenance
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

This study presents an open-source toolkit to address critical challenges in preprocessing data for self-supervised learning (SSL) for 3D medical imaging, focusing on data privacy and computational efficiency. The toolkit comprises two main components: a segmentation network that delineates foreground regions to optimize data sampling and thus reduce training time, and a segmentation network that identifies anonymized regions, preventing erroneous supervision in reconstruction-based SSL methods. Experimental results demonstrate high robustness, with mean Dice scores exceeding 98.5 across all anonymization methods and surpassing 99.5 for foreground segmentation tasks, highlighting the efficacy of the toolkit in supporting SSL applications in 3D medical imaging for both CT and MRI images. The weights and code is available at https://github.com/MIC-DKFZ/Foreground-and-Anonymization-Area-Segmentation.

Problem

• 3D Medical Imaging
• Self-supervised Learning
• Privacy Protection

Innovation

• Privacy-Preserving Processing
• Self-Learning Method
• Efficient Region Identification