Medical image restoration improved by unifying common features and differences
UniH$^3$: Unifying Hierarchical Homogeneity and Heterogeneity for All-in-One Medical Image Restoration
Computer Vision and Pattern Recognition
Summary
Restoring medical images is hard because images can vary a lot by type and damage. The authors created a method called UniH3 that learns what all medical images share in common and what makes them different. Their system uses this knowledge to fix images better and handle many kinds of damage all at once. Tests show UniH3 works better than previous methods on big collections of medical images.
What this means in practice
- •For medical imaging software developers: Integrate a unified model to restore medical images across various scanners and damage types with improved accuracy and efficiency.
- •For radiology workflow engineers: Deploy a universal restoration system that reduces the need for separate models per modality or artifact, streamlining image quality improvement in clinical workflows.
Authors
Zhiwen Yang, Jiayin Li, Chengyu Liu, Hui Zhang, Bingzheng Wei, Yan Xu
Abstract
All-in-One medical image restoration (MedIR) aims to address diverse tasks across modalities and degradation types using a single universal model. Existing methods typically prioritize modeling inter-task heterogeneity (e.g., distinct data distributions and degradation types). However, they largely neglect the inherent homogeneity present in medical images, such as widely shared anatomical structures within and across modalities, which can be leveraged to ease model training and improve generalization. To this end, we propose UniH3, a novel framework that Unifies Hierarchical Homogeneity and Heterogeneity for all-in-one medical image restoration. Specifically, to comprehensively exploit homogeneity, we introduce a Hierarchical Homogeneity Memory (H2M) module that progressively distills intra- and inter-task homogeneity priors from high-quality images during training, and adaptively retrieves the most relevant priors tailored to the input for guided restoration. These retrieved priors are then injected into the restoration pipeline via an efficient Homogeneity-Guided Attention (HGA) mechanism. Furthermore, to comprehensively address heterogeneity, we design a Hierarchical Heterogeneity Balancer (H2B) that mitigates both inter- and intra-task conflicts during optimization, facilitating balanced and effective multi-task learning. Extensive experiments on two large-scale benchmarks, MedIR-2D-500K and MedIR-3D-3K, demonstrate that UniH3 achieves state-of-the-art performance on both all-in-one and single-task medical image restoration. We hope this work establishes a strong benchmark and advances the development of general-purpose medical image restoration models. Code is available at https://github.com/Yaziwel/UniH3.