Registration-Grounded Spectral Fusion for Unregistered WLI/NBI Endoscopic Lesion Segmentation

2026-07-28

AI Summary

This work addresses the challenge of spatial misalignment between white-light imaging (WLI) and narrow-band imaging (NBI) in endoscopic applications, which arises from viewpoint variations, tissue deformation, and sequential acquisition, leading to erroneous fusion and degraded lesion boundary segmentation. To overcome this, the authors propose a reliability-aware complex-domain fusion framework that uniquely models WLI as magnitude and NBI as phase within a complex representation, enabling role-distinct multimodal integration. The method establishes cross-modal feature correspondence through topological regularization and incorporates a reliability estimation mechanism to selectively suppress unreliable interactions in misaligned regions. Evaluated on multiple endoscopic datasets, this end-to-end approach significantly improves segmentation accuracy, with ablation studies confirming the effectiveness of both the modality-specific role design and the reliability-guided fusion strategy.

Abstract

White-light imaging (WLI) and narrow-band imaging (NBI) provide complementary views of endoscopic lesions, but their paired observations are often spatially misaligned due to viewpoint changes, tissue deformation, and sequential handheld acquisition. This makes direct WLI/NBI fusion prone to mixing non-corresponding regions and may even degrade segmentation around lesion boundaries. To address this problem, we propose a reliability-aware complex-domain fusion framework for paired-but-unregistered WLI/NBI lesion segmentation. The framework first establishes topology-regularized feature correspondence and further estimates where the cross-modal correspondence is reliable. Guided by this reliability, the model selectively fuses WLI and NBI features in a learnable complex representation. In this representation, WLI-derived cues mainly provide appearance-related magnitude responses, while NBI-derived cues provide structure-sensitive phase responses. Unlike conventional real-valued or symmetric multimodal fusion, the proposed method explicitly models the different roles of WLI and NBI and suppresses unreliable cross-modal interaction in locally mismatched regions. Experiments on paired WLI/NBI endoscopic datasets show that the proposed reliability-aware registration grounding and complex-domain fusion consistently improve lesion segmentation performance. Role-reversal and module ablation studies further validate the necessity of both the modality-role design and reliability-guided cross-modal interaction.

Problem

• unregistered
• WLI/NBI fusion
• lesion segmentation
• spatial misalignment
• endoscopic imaging

Innovation

• reliability-aware fusion
• complex-domain representation
• unregistered multimodal segmentation
• modality-role modeling
• endoscopic lesion segmentation