| dc.contributor | 資訊系 | |
| dc.creator (作者) | 廖文宏 | |
| dc.creator (作者) | Liao, Wen-Hung;Cheng, Pin-Chieh | |
| dc.date (日期) | 2025-08 | |
| dc.date.accessioned | 3-Oct-2025 09:53:31 (UTC+8) | - |
| dc.date.available | 3-Oct-2025 09:53:31 (UTC+8) | - |
| dc.date.issued (上傳時間) | 3-Oct-2025 09:53:31 (UTC+8) | - |
| dc.identifier.uri (URI) | https://nccur.lib.nccu.edu.tw/handle/140.119/159776 | - |
| dc.description.abstract (摘要) | Overhead fisheye cameras offer broad spatial coverage, making them suitable for surveillance in public spaces such as libraries. However, their severe image distortion and the scarcity of publicly available, privacy-compliant datasets hinder effective object detection. This study addresses these challenges through a dual strategy: augmenting data using text-to-image generative models and correcting fisheye distortion via calibrated intrinsic camera parameters. This approach enables robust training on enriched datasets while mitigating geometric artifacts. Experimental results show substantial performance gains over the YOLOv8 baseline, with mAP@0.5 improving from 0.246 to 0.688 and mAP@0.5:0.95 from 0.122 to 0.518. Detection of small objects—such as beverages—improved markedly, with mAP@0.5 rising from 0.507 to 0.795. Furthermore, combining synthetic and real data in training not only enhances generalization but also improves model robustness under challenging visual conditions. | |
| dc.format.extent | 111 bytes | - |
| dc.format.mimetype | text/html | - |
| dc.relation (關聯) | Proceedings of the 21st IEEE International Conference on Advanced Visual and Signal-Based Systems, IEEE Signal Processing Society | |
| dc.title (題名) | Enhancing Fisheye Lens Object Detection with Generative Data Augmentation | |
| dc.type (資料類型) | conference | |
| dc.identifier.doi (DOI) | 10.1109/AVSS65446.2025.11149842 | |
| dc.doi.uri (DOI) | https://doi.org/10.1109/AVSS65446.2025.11149842 | |
