Hyperbolic Cycle Alignment for Infrared-Visible Image Fusion

2025-07-31

AI Summary

Cross-modal registration of infrared and visible-light images remains challenging due to large modality discrepancies, which hinder conventional Euclidean-space methods from effectively modeling misaligned structures—leading to limited accuracy and robustness. To address this, we propose the Hyperbolic Cycle Alignment Network (HCAN), the first image registration framework operating in hyperbolic space. HCAN introduces a dual-path cyclic architecture comprising forward alignment and backward reconstruction. We further design the Hierarchical Hyperbolic Contrastive Alignment (H²CA) module, which jointly enforces hierarchical contrastive learning and geometric consistency constraints to enhance cross-modal representation alignment and registration sensitivity. Extensive experiments on multiple misaligned cross-modal datasets demonstrate that HCAN significantly outperforms state-of-the-art methods. Both qualitative and quantitative evaluations confirm its superior registration accuracy, fused image quality, and generalization capability across diverse scenarios.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Web applications in cross-disciplinary domains and verticals such as mixed reality, smart cities, and digital health

Abstract

Image fusion synthesizes complementary information from multiple sources, mitigating the inherent limitations of unimodal imaging systems. Accurate image registration is essential for effective multi-source data fusion. However, existing registration methods, often based on image translation in Euclidean space, fail to handle cross-modal misalignment effectively, resulting in suboptimal alignment and fusion quality. To overcome this limitation, we explore image alignment in non-Euclidean space and propose a Hyperbolic Cycle Alignment Network (Hy-CycleAlign). To the best of our knowledge, Hy-CycleAlign is the first image registration method based on hyperbolic space. It introduces a dual-path cross-modal cyclic registration framework, in which a forward registration network aligns cross-modal inputs, while a backward registration network reconstructs the original image, forming a closed-loop registration structure with geometric consistency. Additionally, we design a Hyperbolic Hierarchy Contrastive Alignment (H$^{2}$CA) module, which maps images into hyperbolic space and imposes registration constraints, effectively reducing interference caused by modality discrepancies. We further analyze image registration in both Euclidean and hyperbolic spaces, demonstrating that hyperbolic space enables more sensitive and effective multi-modal image registration. Extensive experiments on misaligned multi-modal images demonstrate that our method significantly outperforms existing approaches in both image alignment and fusion. Our code will be publicly available.

Problem

• Addresses cross-modal misalignment in image registration
• Proposes hyperbolic space for better multi-modal alignment
• Improves fusion quality via geometric consistency constraints

Innovation

• Hyperbolic Cycle Alignment Network for registration
• Dual-path cross-modal cyclic registration framework
• Hyperbolic Hierarchy Contrastive Alignment module