EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

YOLOv4‐dense: A smaller and faster YOLOv4 for real‐time edge‐device based object detection in traffic scene

Yue Jiang, Wenjing Li, Jun Zhang, Fang Li, Zhongcheng Wu

IET Image Processing · 2023-02 · 17 (2) · 570-580

DOI: 10.1049/ipr2.12656 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IET Image Processing, online erschienen am 08.11.2022. Verfasst von Yue Jiang und 4 weiteren Autor:innen. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: High Magnetic Field Laboratory, Chinese Academy of Sciences.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Edge‐device‐based object detection is crucial in many real‐world applications, such as self‐driving cars, ADAS, driver behavior analysis. Although deep learning (DL) has become the de‐facto approach for object detection, the limited computing resources of embedded devices and the large model size of current DL‐based methods increase the difficulty of real‐time object detection on edge devices. To overcome these difficulties, in this work a novel YOLOv4‐dense model is proposed to detect objects in an accurate, fast manner, which is built on top of the YOLOv4 framework but with substantial improvements. More specifically, lots of CSP layers are pruned since it will decrease inference speed. And to address the losing small objects problem, a dense block is introduced. In addition, a lightweight two‐stream YOLO head is also designed to further reduce the computational complexity of the model. Experimental results on NVIDIA JETSON TX2 embedded platform demonstrate that YOLOv4‐dense can achieve a higher accuracy, faster speed with smaller model size. For instance, on the KITTI dataset, YOLOv4‐dense obtains 84.3% mAP and 22.6 FPS with only 20.3 M parameters, surpassing the state‐of‐the‐art models with comparable parameter budget such as YOLOv3‐tiny, YOLOv4‐tiny, PP‐YOLO‐tiny by a large margin.

Förderung

  • High Magnetic Field Laboratory, Chinese Academy of Sciences

Zitieren

APA

Jiang, Y., Li, W., Zhang, J., Li, F., & Wu, Z. (2023). YOLOv4‐dense: A smaller and faster YOLOv4 for real‐time edge‐device based object detection in traffic scene. IET Image Processing, 17(2), 570-580. https://doi.org/10.1049/ipr2.12656

Vancouver

Jiang Y, Li W, Zhang J, Li F, Wu Z. YOLOv4‐dense: A smaller and faster YOLOv4 for real‐time edge‐device based object detection in traffic scene. IET Image Processing. 2023;17(2):570-580. doi:10.1049/ipr2.12656

BibTeX
@article{Jiang2023,
  author = {Jiang, Yue and Li, Wenjing and Zhang, Jun and Li, Fang and Wu, Zhongcheng},
  title = {YOLOv4‐dense: A smaller and faster YOLOv4 for real‐time edge‐device based object detection in traffic scene},
  journal = {IET Image Processing},
  year = {2023},
  volume = {17},
  number = {2},
  pages = {570--580},
  doi = {10.1049/ipr2.12656}
}
Kostenlos starten →