Problem
Research questions and friction points this paper is trying to address.
Retinal Image Registration
Modality-Invariant
Cross-Modal Registration
Generalizability
Modality-Invariant Coarse-to-Fine Retinal Image Registration
🤖 AI Summary
This study addresses the limitation of existing retinal image registration methods that rely on specific modalities and lack cross-modality generalizability. We propose a modality-invariant, two-stage registration framework to overcome these challenges. The approach first leverages universal vessel segmentation to drive sparse matching for coarse alignment, followed by fine registration utilizing a novel mi-raft optical flow network. Experimental results demonstrate that this framework exhibits strong robustness across diverse imaging modality combinations, effectively eliminating the dependency on fixed modalities inherent in traditional approaches. Furthermore, the proposed method achieves registration accuracy significantly superior to current state-of-the-art techniques. Consequently, this work provides a generalized solution for cross-modal retinal image analysis, advancing the field beyond modality-specific constraints.
📝 Abstract
Retinal image registration is essential for ophthalmic diagnosis, longitudinal disease monitoring, and multimodal retinal image analysis. Existing retinal registration methods are typically modality-dependent: they are designed or optimized either for a single imaging modality in mono-modal registration or for a fixed pair of modalities in cross-modal registration. This limits their flexibility and applicability in practical scenarios involving diverse retinal imaging modalities and different combinations of them. In this work, we propose a generalizable two-stage, modality-invariant framework for retinal image registration. First, we introduce a sparse feature-matching model driven by a universal retinal vessel segmentation to achieve robust coarse global alignment across modalities. Second, we develop a modality-invariant optical flow estimation network, termed MI-RAFT, to refine the alignment through dense local registration. Extensive experiments demonstrate that the proposed method can handle diverse combinations of commonly used retinal imaging modalities, exhibiting strong modality invariance while outperforming state-of-the-art modality-dependent registration methods.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Modality-Invariant Registration
MI-RAFT
Coarse-to-Fine Framework
Universal Vessel Segmentation
Cross-Modal Alignment
🔎 Similar Papers
No similar papers found.
