EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

GFLNet: A lightweight real-time detector with knowledge-guided compression in remote sensing images

Jing Zhang, Mas Rina Mustaffa iD, Fatimah Khalid, Zainal Abdul Kahar, Tao Zhang

PLOS One · 30.10.2026 · 21 (10) · e0361806

DOI: 10.1371/journal.pone.0361806 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 30.10.2026. Verfasst von Jing Zhang und 4 weiteren Autor:innen. Beteiligte Einrichtung: Faculty of Computer Science and Information Technology, Universiti Putra Malaysia, Serdang, Malaysia. Hauptthemen laut MEDLINE: Remote Sensing Technology, Data Compression, Image Processing, Computer-Assisted, Compression Algorithms. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Remote sensing images usually exhibits densely distributed targets and scale disparities. In such scenarios, it is still challenging to preserve high detection precision while controlling costs and inference latency. To address this issue, we propose GFLNet, a lightweight single-stage detector, which is based on collaborative feature modeling and knowledge-guided compression. Specifically, Ghost-Shuffle dual fusion convolution (GSD-Fuse) is designed to achieve efficient spatial-channel interaction while reducing computational cost; an adaptive receptive-field selection module inspired by LSKNet is introduced into the deeper backbone and neck stages to adaptively adjust contextual responses for small-object and multi-scale detection; finally, LB-Compress combines layer-adaptive magnitude pruning with Bidirectional Cross-task Knowledge Distillation (BCKD) to form a compact network suitable for practical deployment. Evaluated on the DIOR and UCAS-AOD datasets, GFLNet improves mAP@0.5 over the baseline by 3.9% and 3.4%, respectively. Moreover, the model contains only 2.35M parameters, with a calculation of 5.6 GFLOPs, and an inference speed of 261 FPS on the GPU, thereby delivering a favorable accuracy-efficiency synergy.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Remote Sensing TechnologyData CompressionImage Processing, Computer-AssistedCompression AlgorithmsAlgorithms

Autor:innen und Institutionen

  • Faculty of Computer Science and Information Technology, Universiti Putra Malaysia, Serdang, Malaysia.
  • Department of General Education, Anhui Vocational College of City Management, Hefei, China.
  • Whole vehicle development-Material technology, Volkswagen China Technology Company, Hefei, Anhui, China.

ORCID-Profile: Mas Rina Mustaffa

Zitieren

APA

Zhang, J., Mustaffa, M. R., Khalid, F., Kahar, Z. A., & Zhang, T. (2026). GFLNet: A lightweight real-time detector with knowledge-guided compression in remote sensing images. PLOS One, 21(10), e0361806. https://doi.org/10.1371/journal.pone.0361806

Vancouver

Zhang J, Mustaffa MR, Khalid F, Kahar ZA, Zhang T. GFLNet: A lightweight real-time detector with knowledge-guided compression in remote sensing images. PLOS One. 2026;21(10):e0361806. doi:10.1371/journal.pone.0361806

BibTeX
@article{Zhang2026,
  author = {Zhang, Jing and Mustaffa, Mas Rina and Khalid, Fatimah and Kahar, Zainal Abdul and Zhang, Tao},
  title = {GFLNet: A lightweight real-time detector with knowledge-guided compression in remote sensing images},
  journal = {PLOS One},
  year = {2026},
  volume = {21},
  number = {10},
  pages = {e0361806},
  doi = {10.1371/journal.pone.0361806}
}
Kostenlos starten →