EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Parameter Efficient Adaptation of Vision Large Language Models for Fisheye Object Detection

Amine Bechar, Abbes Amira, Fatih Kurugollu, Adel Oulefki, Munkhjargal Gochoo, Fady Alnajjar

IEEE Open Journal of Signal Processing · 2026 · 7 · 746-752

DOI: 10.1109/ojsp.2026.3706022 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Open Journal of Signal Processing, 2026. Verfasst von Amine Bechar und 5 weiteren Autor:innen. Die Arbeit nennt 17 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: University of Sharjah, United Arab Emirates University, United Arab Emirates.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Fisheye cameras are used for coverage in transportation systems and vehicles, yet distortions pose challenges to object detection algorithms. In this paper, Radial-Aware LoRA (RA-LoRA) is proposed, a distortion-aware parameter-efficient adaptation method that modulates low-rank updates based on radial position to address the spatially varying nature of fisheye distortion. RA-LoRA fine-tuning of Florence-2-large with only 0.5% additional parameters achieves a 181% relative improvement in mAP@50 over zero-shot inference, outperforming fully trained CNN detectors in fisheye imagery. Detection performance is assessed in all five object classes of the FishEye8K dataset (bike, bus, car, pedestrian, and truck). The results show that Florence-2-large with RA-LoRA achieves a mean average precision (mAP) of 0.724, outperforming YOLO26x (mAP: 0.581) and YOLO26 l (mAP: 0.605). Although the model requires approximately 4.5× more memory than YOLO26, the RA-LoRA adapters represent only 0.5% of the total parameters and can be merged into the base model at inference time with no additional overhead. This makes the proposed approach particularly suitable for server-side processing, offline traffic analysis, post-incident auditing, and high-accuracy surveillance applications where detection reliability outweighs the need for real-time embedded deployment.

Förderung

  • University of Sharjah
  • United Arab Emirates University
  • United Arab Emirates

Zitieren

APA

Bechar, A., Amira, A., Kurugollu, F., Oulefki, A., Gochoo, M., & Alnajjar, F. (2026). Parameter Efficient Adaptation of Vision Large Language Models for Fisheye Object Detection. IEEE Open Journal of Signal Processing, 7, 746-752. https://doi.org/10.1109/ojsp.2026.3706022

Vancouver

Bechar A, Amira A, Kurugollu F, Oulefki A, Gochoo M, Alnajjar F. Parameter Efficient Adaptation of Vision Large Language Models for Fisheye Object Detection. IEEE Open Journal of Signal Processing. 2026;7:746-752. doi:10.1109/ojsp.2026.3706022

BibTeX
@article{Bechar2026,
  author = {Bechar, Amine and Amira, Abbes and Kurugollu, Fatih and Oulefki, Adel and Gochoo, Munkhjargal and Alnajjar, Fady},
  title = {Parameter Efficient Adaptation of Vision Large Language Models for Fisheye Object Detection},
  journal = {IEEE Open Journal of Signal Processing},
  year = {2026},
  volume = {7},
  pages = {746--752},
  doi = {10.1109/ojsp.2026.3706022}
}
Kostenlos starten →