Efficient and robust 3D blind harmonization for large domain gaps

2025-04-30

AI Summary

Existing 3D blind harmonization methods lack robustness under large cross-scanner domain shifts, slice-wise heterogeneity, degraded image quality, and unavailability of source-domain data—common clinical constraints in MRI. Method: We propose the first edge-to-image 3D blind harmonization paradigm. It employs a multi-scale voxel-block training strategy to jointly capture local and global anatomical priors; integrates an edge-guided reconstruction and hallucination-suppression refinement module; and synergistically combines a 3D correction flow model with an edge-driven source-domain mapping network—all operating end-to-end using target-domain data only. Contribution/Results: Our method achieves state-of-the-art performance on multi-vendor MRI datasets. Downstream tasks—brain tissue segmentation and chronological age prediction—show significant improvements: Dice score increases by 12.7% and coefficient of determination (R²) improves by 8.3%. The framework markedly enhances clinical generalizability and practical deployability without requiring source-domain access.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Blind harmonization has emerged as a promising technique for MR image harmonization to achieve scale-invariant representations, requiring only target domain data (i.e., no source domain data necessary). However, existing methods face limitations such as inter-slice heterogeneity in 3D, moderate image quality, and limited performance for a large domain gap. To address these challenges, we introduce BlindHarmonyDiff, a novel blind 3D harmonization framework that leverages an edge-to-image model tailored specifically to harmonization. Our framework employs a 3D rectified flow trained on target domain images to reconstruct the original image from an edge map, then yielding a harmonized image from the edge of a source domain image. We propose multi-stride patch training for efficient 3D training and a refinement module for robust inference by suppressing hallucination. Extensive experiments demonstrate that BlindHarmonyDiff outperforms prior arts by harmonizing diverse source domain images to the target domain, achieving higher correspondence to the target domain characteristics. Downstream task-based quality assessments such as tissue segmentation and age prediction on diverse MR scanners further confirm the effectiveness of our approach and demonstrate the capability of our robust and generalizable blind harmonization.

Problem

• Addressing inter-slice heterogeneity in 3D blind harmonization
• Improving image quality for large domain gaps
• Enhancing robustness and generalizability in MR harmonization

Innovation

• 3D rectified flow for image reconstruction
• Edge-to-image model for harmonization
• Multi-stride patch training for efficiency