PuSH - Publication Server of Helmholtz Zentrum München

Liu, Y.* ; Ma, R. ; Navab, N.* ; Cui, J. ; Peng, T.

CURVFUSE: Field-Curvature-Aware Multi-Focus Fusion for Bioluminescence Imaging.

In: (23rd IEEE International Symposium on Biomedical Imaging, ISBI 2026, 8-11 April 2026, London). 2026. (Proceedings International Symposium on Biomedical Imaging ; 2026-April)
DOI
Field curvature, a common phenomenon that causes distinct regions of a specimen to come into focus at different axial depths, limits the usable field of view in optical microscopy. Although recent image fusion approaches locally compare the image quality of two inputs and identify in-focus pixels, they often produce spatially inconsistent focus measures considering the sparse cellular structures. Moreover, the long exposure time of methods such as bioluminescence imaging, an emerging field, makes the specimen itself nonstationary, as cells may move, deform, or die, which further complicates the fusion. Here, we present CurvFuse, a field-curvature-aware fusion framework that explicitly estimates a spatially coherent fusion boundary guided by the image formation prior of depth-dependent focal surfaces. Morphology-guided operations and structural constraints are further integrated to ensure boundary robustness against local deformation and to maintain structural fidelity throughout the fusion process. To the best of our knowledge, CurvFuse is the first bioluminescence fusion method that enables the expansion of the usable field of view without inducing artifacts. All implementations are made publicly available at https://github.com/penglab/CurvFuse.
Altmetric
Additional Metrics?
[➜Log in]
Edit extra informations Login
Publication type Article: Conference contribution
Keywords Bioluminescence Imaging ; Field Curvature ; Monochromatic Aberration ; Multi-focus Fusion
ISSN (print) / ISBN 1945-7928
e-ISSN 1945-8452
Conference Title 23rd IEEE International Symposium on Biomedical Imaging, ISBI 2026
Conference Date 8-11 April 2026
Conference Location London
Quellenangaben Volume: 2026-April Issue: , Pages: , Article Number: , Supplement: ,
Institute(s) Helmholtz Artifical Intelligence Cooperation Unit (HAICU)
Institute of Radiation Medicine (IRM)
Helmholtz Pioneer Campus (HPC)