A Review on Underwater Image Enhancement Models, Datasets and Metrics
Computer Life · 2024
DOI: 10.54097/5nbdy596 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computer Life, 2024. Verfasst von Lingyan Kong und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
Exploration of the underwater world is still the direction we are heading. Underwater imaging remains challenging due to blurred visibility, color deviation, and low contrast. Underwater image enhancement (UIE) represents a fundamental yet critical research challenge in the field of computer vision. Despite continuous advancements in hardware and algorithmic methodologies, there remains a lack of comprehensive summaries in this domain. To address this, we provide an overview of the research progress in UIE from the following perspectives. First, we introduce the three mainstream categories of UIE algorithms, along with the construction of datasets, including paired and unpaired datasets. Second, we conduct model training and evaluate the performance on a unified dataset, presenting results from both quantitative and qualitative perspectives. Finally, by utilizing the enhanced datasets for object detection tasks, we observe that the evaluation metrics of image enhancement and object detection do not exhibit a positive correlation.
Zitieren
APA
Kong, L., & Li, Z. (2024). A Review on Underwater Image Enhancement Models, Datasets and Metrics. Computer Life. https://doi.org/10.54097/5nbdy596
Vancouver
Kong L, Li Z. A Review on Underwater Image Enhancement Models, Datasets and Metrics. Computer Life. 2024. doi:10.54097/5nbdy596
BibTeX
@article{Kong2024,
author = {Kong, Lingyan and Li, Zhanying},
title = {A Review on Underwater Image Enhancement Models, Datasets and Metrics},
journal = {Computer Life},
year = {2024},
doi = {10.54097/5nbdy596}
}
EUVIMED