Downsampling Shuffle Transformer for Underwater Image Enhancement
IEEE Signal Processing Letters · 2026 · 33 · 1976-1980
DOI: 10.1109/lsp.2026.3690576 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Wanying Zhang und 3 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
Underwater scenes, characterized by light absorption and scattering, frequently result in color distortion and low contrast in captured images. Current deep learning-based methods (like CNNs and Transformers) have advanced underwater image enhancement, yet window-based partitioning of standard Transformers limits non-local interaction modeling for efficiency. In this paper, we present a novel window-partitioning strategy dubbed downsampling shuffle, enabling the local window Transformer to capture non-local interactions. By pixel-spaced downsampling of original features, the partitioned window retains a near-global receptive field, allowing local window self-attention to model non-local interactions without extra computation costs. The resulting Transformer dubbed DSFormer efficiently processes images while maintaining a global receptive field, which is crucial for underwater image enhancement. Extensive experiments are conducted to verify the superiority of our DSFormer across public datasets.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Zhang, W., Shang, W., Ren, D., & Zuo, W. (2026). Downsampling Shuffle Transformer for Underwater Image Enhancement. IEEE Signal Processing Letters, 33, 1976-1980. https://doi.org/10.1109/lsp.2026.3690576
Vancouver
Zhang W, Shang W, Ren D, Zuo W. Downsampling Shuffle Transformer for Underwater Image Enhancement. IEEE Signal Processing Letters. 2026;33:1976-1980. doi:10.1109/lsp.2026.3690576
BibTeX
@article{Zhang2026,
author = {Zhang, Wanying and Shang, Wei and Ren, Dongwei and Zuo, Wangmeng},
title = {Downsampling Shuffle Transformer for Underwater Image Enhancement},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {1976--1980},
doi = {10.1109/lsp.2026.3690576}
}
EUVIMED