RETRACTED: Underwater image enhancement using multi-task fusion
PLOS ONE · 26.02.2024 · 19 (2) · e0299110
DOI: 10.1371/journal.pone.0299110 ↗Kurz gefasst
Zurückgezogene Publikation in der Zeitschrift PLOS ONE, online erschienen am 26.02.2024. Verfasst von Kaibo Liao und 1 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer, Central China Normal University, Wuhan, Hubei, China. Hauptthemen laut MEDLINE: Algorithms, Image Enhancement. Die Arbeit wurde bisher 32-mal zitiert (Crossref) und nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Underwater images are often scattered due to suspended particles in the water, resulting in light scattering and blocking and reduced visibility and contrast. Color shifts and distortions are also caused by the absorption of different wavelengths of light in the water. This series of problems will make the underwater image quality greatly impaired, resulting in some advanced visual work can not be carried out underwater. In order to solve these problems, this paper proposes an underwater image enhancement method based on multi-task fusion, called MTF. Specifically, we first use linear constraints on the input image to achieve color correction based on the gray world assumption. The corrected image is then used to achieve visibility enhancement using an improved type-II fuzzy set-based algorithm, while the image is contrast enhanced using standard normal distribution probability density function and softplus function. However, in order to obtain more qualitative results, we propose multi-task fusion, in which we solve for similarity, then we obtain fusion weights that guarantee the best features of the image as much as possible from the obtained similarity, and finally we fuse the image with the weights to obtain the output image, and we find that multi-task fusion has excellent image enhancement and restoration capabilities, and also produces visually pleasing results. Extensive qualitative and quantitative evaluations show that MTF method achieves optimal results compared to ten state-of-the-art underwater enhancement algorithms on 2 datasets. Moreover, the method can achieve better results in application tests such as target detection and edge detection.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- School of Computer, Central China Normal University, Wuhan, Hubei, China.
- Hubei Provincial Key Laboratory of Artificial Intelligence and Smart Learning, Central China Normal University, Wuhan, Hubei, China.
ORCID-Profile: Kaibo Liao
Zitieren
APA
Liao, K., & Peng, X. (2024). RETRACTED: Underwater image enhancement using multi-task fusion. PLOS ONE, 19(2), e0299110. https://doi.org/10.1371/journal.pone.0299110
Vancouver
Liao K, Peng X. RETRACTED: Underwater image enhancement using multi-task fusion. PLOS ONE. 2024;19(2):e0299110. doi:10.1371/journal.pone.0299110
BibTeX
@article{Liao2024,
author = {Liao, Kaibo and Peng, Xi},
title = {RETRACTED: Underwater image enhancement using multi-task fusion},
journal = {PLOS ONE},
year = {2024},
volume = {19},
number = {2},
pages = {e0299110},
doi = {10.1371/journal.pone.0299110}
}
EUVIMED