Advancing RGB-IR Image Fusion: Exploiting Hyperbolic Geometry
Infrared and visible image fusion is essential for remote sensing applications, especially for obtaining high-quality imagery of terrestrial environments. Hierarchical feature information is crucial for image fusion as it captures the intricate relationships between different modalities, which are v...
| Prif Awduron: | , , |
|---|---|
| Fformat: | Erthygl |
| Iaith: | English |
| Cyhoeddwyd: |
IEEE
2025-01-01
|
| Cyfres: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Pynciau: | |
| Mynediad Ar-lein: | https://ieeexplore.ieee.org/document/10878430/ |
| _version_ | 1826575444640530432 |
|---|---|
| author | Chenyu Peng Tao Shen Qingwang Wang |
| author_facet | Chenyu Peng Tao Shen Qingwang Wang |
| author_sort | Chenyu Peng |
| collection | DOAJ |
| description | Infrared and visible image fusion is essential for remote sensing applications, especially for obtaining high-quality imagery of terrestrial environments. Hierarchical feature information is crucial for image fusion as it captures the intricate relationships between different modalities, which are vital for producing detailed and accurate composite images. However, most existing methods operate within the confines of Euclidean space, which, due to its inherently “flat” geometric nature, often struggles to effectively measure the similarities and differences between modalities, thus failing to maintain their distinctiveness. Hyperbolic space, with its constant negative curvature, excels at leveraging these hierarchical structures. It can more effectively gauge the similarities and differences between modalities, preserving their distinctiveness. In this study, we propose a novel fusion method for infrared and visible image fusion in hyperbolic space, named HbFNet. We have developed innovative hyperbolic feature extraction modules, including Hyperbolic Invertible Neural Networks and Hyperbolic Lite Transformer blocks, specifically designed to capitalize on the hierarchical nature of features. Our method emerges as a promising solution for enhancing hierarchical information and elevating the quality of fusion. Extensive experiments across three public datasets have demonstrated that our method outperforms most state-of-the-art image fusion techniques. |
| first_indexed | 2025-03-14T13:14:48Z |
| format | Article |
| id | doaj.art-4b51c27e16c94ff9aea1f61a3f754bc7 |
| institution | Directory Open Access Journal |
| issn | 1939-1404 2151-1535 |
| language | English |
| last_indexed | 2025-03-14T13:14:48Z |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-4b51c27e16c94ff9aea1f61a3f754bc72025-03-01T00:00:06ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01186007601610.1109/JSTARS.2025.354030410878430Advancing RGB-IR Image Fusion: Exploiting Hyperbolic GeometryChenyu Peng0https://orcid.org/0009-0005-7431-0647Tao Shen1https://orcid.org/0000-0003-1273-7950Qingwang Wang2https://orcid.org/0000-0001-5820-5357Faculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming, ChinaFaculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming, ChinaFaculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming, ChinaInfrared and visible image fusion is essential for remote sensing applications, especially for obtaining high-quality imagery of terrestrial environments. Hierarchical feature information is crucial for image fusion as it captures the intricate relationships between different modalities, which are vital for producing detailed and accurate composite images. However, most existing methods operate within the confines of Euclidean space, which, due to its inherently “flat” geometric nature, often struggles to effectively measure the similarities and differences between modalities, thus failing to maintain their distinctiveness. Hyperbolic space, with its constant negative curvature, excels at leveraging these hierarchical structures. It can more effectively gauge the similarities and differences between modalities, preserving their distinctiveness. In this study, we propose a novel fusion method for infrared and visible image fusion in hyperbolic space, named HbFNet. We have developed innovative hyperbolic feature extraction modules, including Hyperbolic Invertible Neural Networks and Hyperbolic Lite Transformer blocks, specifically designed to capitalize on the hierarchical nature of features. Our method emerges as a promising solution for enhancing hierarchical information and elevating the quality of fusion. Extensive experiments across three public datasets have demonstrated that our method outperforms most state-of-the-art image fusion techniques.https://ieeexplore.ieee.org/document/10878430/Hyperbolic convolutionhyperbolic spaceimage fusionremote sensing |
| spellingShingle | Chenyu Peng Tao Shen Qingwang Wang Advancing RGB-IR Image Fusion: Exploiting Hyperbolic Geometry IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Hyperbolic convolution hyperbolic space image fusion remote sensing |
| title | Advancing RGB-IR Image Fusion: Exploiting Hyperbolic Geometry |
| title_full | Advancing RGB-IR Image Fusion: Exploiting Hyperbolic Geometry |
| title_fullStr | Advancing RGB-IR Image Fusion: Exploiting Hyperbolic Geometry |
| title_full_unstemmed | Advancing RGB-IR Image Fusion: Exploiting Hyperbolic Geometry |
| title_short | Advancing RGB-IR Image Fusion: Exploiting Hyperbolic Geometry |
| title_sort | advancing rgb ir image fusion exploiting hyperbolic geometry |
| topic | Hyperbolic convolution hyperbolic space image fusion remote sensing |
| url | https://ieeexplore.ieee.org/document/10878430/ |
| work_keys_str_mv | AT chenyupeng advancingrgbirimagefusionexploitinghyperbolicgeometry AT taoshen advancingrgbirimagefusionexploitinghyperbolicgeometry AT qingwangwang advancingrgbirimagefusionexploitinghyperbolicgeometry |