# Efficient and robust 3D blind harmonization for large domain gaps

arXiv: https://arxiv.org/abs/2505.00133 · HTML: https://mlsanity.com/paper/2505.00133/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 54 references that are in the map.
- First posted on arXiv 30 Apr 2025.

## Builds on (most cited first, 24 of 24)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow](https://mlsanity.com/paper/2209.03003/) (2022)
- [Consistency Models](https://mlsanity.com/paper/2303.01469/) (2023)
- [Neural Spline Flows](https://mlsanity.com/paper/1906.04032/) (2019)
- [Learning a Variational Network for Reconstruction of Accelerated MRI Data](https://mlsanity.com/paper/1704.00447/) (2017)
- [Patch Diffusion: Faster and More Data-Efficient Training of Diffusion Models](https://mlsanity.com/paper/2304.12526/) (2023)
- [Solving 3D Inverse Problems using Pre-trained 2D Diffusion Models](https://mlsanity.com/paper/2211.10655/) (2022)
- [Understanding Hallucinations in Diffusion Models through Mode Interpolation](https://mlsanity.com/paper/2406.09358/) (2024)
- [Understanding SSIM](https://mlsanity.com/paper/2006.13846/) (2020)
- [Predicting brain age with deep learning from raw imaging data results in a reliable and heritable biomarker](https://mlsanity.com/paper/1612.02572/) (2016)
- [Improving 3D Imaging with Pre-Trained Perpendicular 2D Diffusion Models](https://mlsanity.com/paper/2303.08440/) (2023)
- [Residual and Plain Convolutional Neural Networks for 3D Brain MRI Classification](https://mlsanity.com/paper/1701.06643/) (2017)
- [Patched Denoising Diffusion Models For High-Resolution Image Synthesis](https://mlsanity.com/paper/2308.01316/) (2023)
- [A deep learning model integrating FCNNs and CRFs for brain tumor segmentation](https://mlsanity.com/paper/1702.04528/) (2017)
- [Memory-Efficient 3D Denoising Diffusion Models for Medical Image Processing](https://mlsanity.com/paper/2303.15288/) (2023)
- [Quantitative Susceptibility Mapping using Deep Neural Network: QSMnet](https://mlsanity.com/paper/1803.05627/) (2018)
- [Tackling Structural Hallucination in Image Translation with Local Diffusion](https://mlsanity.com/paper/2404.05980/) (2024)
- [ImUnity: a generalizable VAE-GAN solution for multicenter MR image harmonization](https://mlsanity.com/paper/2109.06756/) (2021)
- [Deep Generative Models for 3D Medical Image Synthesis](https://mlsanity.com/paper/2410.17664/) (2024)
- [DiffusionBlend: Learning 3D Image Prior through Position-aware Diffusion Score Blending for 3D Computed Tomography Reconstruction](https://mlsanity.com/paper/2406.10211/) (2024)
- [BlindHarmony: "Blind" Harmonization for MR Images via Flow model](https://mlsanity.com/paper/2305.10732/) (2023)
- [Reliable Source Approximation: Source-Free Unsupervised Domain Adaptation for Vestibular Schwannoma MRI Segmentation](https://mlsanity.com/paper/2405.16102/) (2024)
- [Harmonizing Flows: Leveraging normalizing flows for unsupervised and source-free MRI harmonization](https://mlsanity.com/paper/2407.15717/) (2024)

## Abstract (arXiv)

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.