TranSDet: Toward Effective Transfer Learning for Small-Object Detection
Remote Sensing · 12.07.2023 · 15 (14) · 3525
DOI: 10.3390/rs15143525 ↗Kurz gefasst
Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 12.07.2023. Verfasst von Xinkai Xu und 5 weiteren Autor:innen. Die Arbeit wurde bisher 32-mal zitiert (Crossref) und nennt 63 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Key project of the National Nature Science Foundation of China, Science and Technology Major Project of Shanxi Province, China, Beijing Municipal Education Commission Science and Technology Program.
Abstract (Originalsprache)
Small-object detection is a challenging task in computer vision due to the limited training samples and low-quality images. Transfer learning, which transfers the knowledge learned from a large dataset to a small dataset, is a popular method for improving performance on limited data. However, we empirically find that due to the dataset discrepancy, directly transferring the model trained on a general object dataset to small-object datasets obtains inferior performance. In this paper, we propose TranSDet, a novel approach for effective transfer learning for small-object detection. Our method adapts a model trained on a general dataset to a small-object-friendly model by augmenting the training images with diverse smaller resolutions. A dynamic resolution adaptation scheme is employed to ensure consistent performance on various sizes of objects using meta-learning. Additionally, the proposed method introduces two network components, an FPN with shifted feature aggregation and an anchor relation module, which are compatible with transfer learning and effectively improve small-object detection performance. Extensive experiments on the TT100K, BUUISE-MO-Lite, and COCO datasets demonstrate that TranSDet achieves significant improvements compared to existing methods. For example, on the TT100K dataset, TranSDet outperforms the state-of-the-art method by 8.0% in terms of the mean average precision (mAP) for small-object detection. On the BUUISE-MO-Lite dataset, TranSDet improves the detection accuracy of RetinaNet and YOLOv3 by 32.2% and 12.8%, respectively.
Förderung
- Key project of the National Nature Science Foundation of China
- Science and Technology Major Project of Shanxi Province, China
- Beijing Municipal Education Commission Science and Technology Program
- Academic Research Projects of Beijing Union University
Zitieren
APA
Xu, X., Zhang, H., Ma, Y., Liu, K., Bao, H., & Qian, X. (2023). TranSDet: Toward Effective Transfer Learning for Small-Object Detection. Remote Sensing, 15(14), 3525. https://doi.org/10.3390/rs15143525
Vancouver
Xu X, Zhang H, Ma Y, Liu K, Bao H, Qian X. TranSDet: Toward Effective Transfer Learning for Small-Object Detection. Remote Sensing. 2023;15(14):3525. doi:10.3390/rs15143525
BibTeX
@article{Xu2023,
author = {Xu, Xinkai and Zhang, Hailan and Ma, Yan and Liu, Kang and Bao, Hong and Qian, Xu},
title = {TranSDet: Toward Effective Transfer Learning for Small-Object Detection},
journal = {Remote Sensing},
year = {2023},
volume = {15},
number = {14},
pages = {3525},
doi = {10.3390/rs15143525}
}
EUVIMED