| dc.contributor | 資訊系 | |
| dc.creator (作者) | 彭彥璁 | |
| dc.creator (作者) | Peng, Yan-Tsung;Su, Li;Chen, Zihao;Suarez, Fatima Lois | |
| dc.date (日期) | 2025-07 | |
| dc.date.accessioned | 20-Jan-2026 13:18:57 (UTC+8) | - |
| dc.date.available | 20-Jan-2026 13:18:57 (UTC+8) | - |
| dc.date.issued (上傳時間) | 20-Jan-2026 13:18:57 (UTC+8) | - |
| dc.identifier.uri (URI) | https://ah.lib.nccu.edu.tw/item?item_id=180739 | - |
| dc.description.abstract (摘要) | Images captured in underwater environments often suffer from various degradations, such as color cast, poor lighting, low contrast, and turbidity. Consequently, underwater image enhancement gained popularity. However, obtaining paired real underwater image data remains a challenge, limiting the generalization of existing models. To address this, we propose a data augmentation technique inspired by domain adaptation that leverages domain-specific features from the target domain to augment the training dataset, improving model generalization on unseen data in various real-world target domains without ground truth. Comprehensive experiments demonstrated performance improvements across multiple datasets. | |
| dc.format.extent | 118 bytes | - |
| dc.format.mimetype | text/html | - |
| dc.relation (關聯) | 2025 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-Taiwan), Institute of Electrical and Electronics Engineers (IEEE) | |
| dc.subject (關鍵詞) | underwater image enhancement; data augmentation; domain adaptation; image formation model | |
| dc.title (題名) | Underwater Image Restoration based on Domain-adaptive Data Augmentation | |
| dc.type (資料類型) | conference | |
| dc.identifier.doi (DOI) | 10.1109/ICCE-Taiwan66881.2025.11207785 | |
| dc.doi.uri (DOI) | https://doi.org/10.1109/ICCE-Taiwan66881.2025.11207785 | |
