SinGAN-Labeler: An Enhanced SinGAN for Generating Marine Oil Spill SAR Images with Labels

Deep learning-based SAR oil spill detection faces significant challenges due to limited labeled training data. To address this, we propose SinGAN-Labeler, an enhanced framework that generates high-quality synthetic SAR oil spill images and their labels from minimal input. The model integrates an ada...

Full description

Saved in:
Bibliographic Details
Main Authors: Bin Wang, Lei Chen, Dongmei Song, Weimin Chen, Jintao Yu
Format: Article
Language:English
Published: MDPI AG 2025-02-01
Series:Journal of Marine Science and Engineering
Subjects:
Online Access:https://www.mdpi.com/2077-1312/13/3/422
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849341676649185280
author Bin Wang
Lei Chen
Dongmei Song
Weimin Chen
Jintao Yu
author_facet Bin Wang
Lei Chen
Dongmei Song
Weimin Chen
Jintao Yu
author_sort Bin Wang
collection DOAJ
description Deep learning-based SAR oil spill detection faces significant challenges due to limited labeled training data. To address this, we propose SinGAN-Labeler, an enhanced framework that generates high-quality synthetic SAR oil spill images and their labels from minimal input. The model integrates an adaptive module to automate scale parameter optimization, accelerating training, and a hybrid attention module combining spatial, channel, and global contextual mechanisms to enhance feature extraction. By leveraging multi-scale training with diverse receptive fields, the generated images retain critical structural details while ensuring diversity. Experiments demonstrate that detection models trained on synthetic data achieve performance comparable to those using real images. Notably, expanding data sets by fivefold (from 5, 10, and 15 baseline images) improves the UNet++ model’s IoU by 78.2%, 58.5%, and 22.5%, respectively. These results validate SinGAN-Labeler’s capability to mitigate data scarcity and enhance oil spill detection accuracy, particularly under extreme sample limitations.
format Article
id doaj-art-a7adf99c8b964bf19006fc6095f6aa06
institution Kabale University
issn 2077-1312
language English
publishDate 2025-02-01
publisher MDPI AG
record_format Article
series Journal of Marine Science and Engineering
spelling doaj-art-a7adf99c8b964bf19006fc6095f6aa062025-08-20T03:43:34ZengMDPI AGJournal of Marine Science and Engineering2077-13122025-02-0113342210.3390/jmse13030422SinGAN-Labeler: An Enhanced SinGAN for Generating Marine Oil Spill SAR Images with LabelsBin Wang0Lei Chen1Dongmei Song2Weimin Chen3Jintao Yu4The College of Oceanography and Space Informatics, China University of Petroleum (East China), Qingdao 266404, ChinaThe College of Oceanography and Space Informatics, China University of Petroleum (East China), Qingdao 266404, ChinaThe College of Oceanography and Space Informatics, China University of Petroleum (East China), Qingdao 266404, ChinaThe College of Oceanography and Space Informatics, China University of Petroleum (East China), Qingdao 266404, ChinaThe College of Oceanography and Space Informatics, China University of Petroleum (East China), Qingdao 266404, ChinaDeep learning-based SAR oil spill detection faces significant challenges due to limited labeled training data. To address this, we propose SinGAN-Labeler, an enhanced framework that generates high-quality synthetic SAR oil spill images and their labels from minimal input. The model integrates an adaptive module to automate scale parameter optimization, accelerating training, and a hybrid attention module combining spatial, channel, and global contextual mechanisms to enhance feature extraction. By leveraging multi-scale training with diverse receptive fields, the generated images retain critical structural details while ensuring diversity. Experiments demonstrate that detection models trained on synthetic data achieve performance comparable to those using real images. Notably, expanding data sets by fivefold (from 5, 10, and 15 baseline images) improves the UNet++ model’s IoU by 78.2%, 58.5%, and 22.5%, respectively. These results validate SinGAN-Labeler’s capability to mitigate data scarcity and enhance oil spill detection accuracy, particularly under extreme sample limitations.https://www.mdpi.com/2077-1312/13/3/422marine oil spillgenerative adversarial network (GAN)hybrid attention mechanismSAR image expansion
spellingShingle Bin Wang
Lei Chen
Dongmei Song
Weimin Chen
Jintao Yu
SinGAN-Labeler: An Enhanced SinGAN for Generating Marine Oil Spill SAR Images with Labels
Journal of Marine Science and Engineering
marine oil spill
generative adversarial network (GAN)
hybrid attention mechanism
SAR image expansion
title SinGAN-Labeler: An Enhanced SinGAN for Generating Marine Oil Spill SAR Images with Labels
title_full SinGAN-Labeler: An Enhanced SinGAN for Generating Marine Oil Spill SAR Images with Labels
title_fullStr SinGAN-Labeler: An Enhanced SinGAN for Generating Marine Oil Spill SAR Images with Labels
title_full_unstemmed SinGAN-Labeler: An Enhanced SinGAN for Generating Marine Oil Spill SAR Images with Labels
title_short SinGAN-Labeler: An Enhanced SinGAN for Generating Marine Oil Spill SAR Images with Labels
title_sort singan labeler an enhanced singan for generating marine oil spill sar images with labels
topic marine oil spill
generative adversarial network (GAN)
hybrid attention mechanism
SAR image expansion
url https://www.mdpi.com/2077-1312/13/3/422
work_keys_str_mv AT binwang singanlabeleranenhancedsinganforgeneratingmarineoilspillsarimageswithlabels
AT leichen singanlabeleranenhancedsinganforgeneratingmarineoilspillsarimageswithlabels
AT dongmeisong singanlabeleranenhancedsinganforgeneratingmarineoilspillsarimageswithlabels
AT weiminchen singanlabeleranenhancedsinganforgeneratingmarineoilspillsarimageswithlabels
AT jintaoyu singanlabeleranenhancedsinganforgeneratingmarineoilspillsarimageswithlabels