Citation

Ellen DAR, Kristalina P, Hadi MZS, Patriarso A. IEEE Xplore 2023; 569-574.

Copyright

(Copyright © 2023, Institute of Electrical and Electronics Engineers (IEEE))

DOI

10.1109/IES59143.2023.10242589

PMID

unavailable

Abstract

2023 International Electronics Symposium (IES)

The search for drowning victims can be a time-consuming and challenging task, often hindered by limited equipment in the field. However, advancements in technology, particularly underwater drone technology and deep learning methods for object detection, present promising solutions. YOLOv5 (You Only Look Once Version 5) is a cutting-edge object detection model known for its high detection speed and accuracy. In this study, the researchers leveraged the capabilities of YOLOv5 to enhance the efficiency of searching for submerged victims. The YOLOv5s configuration model was utilized, and the model was trained for 400 epochs. The results demonstrated impressive performance, with an accuracy value of 89%, precision of 100%, and recall of 95%.

Language: en

Keywords

Adaptation models; Computational modeling; deep learning; Deep learning; detection of drowning humans; human detection; Object detection; Real-time systems; Rivers; Training; underwater drone; YOLOv5