EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep Learning with Attention Mechanisms for Road Weather Detection

Madiha Samo, Jimiama Mosima Mafeni Mase iD, Grazziela Figueredo iD

Sensors · 2023

DOI: 10.3390/s23020798 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, 2023. Verfasst von Madiha Samo und 2 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer Science, University of Nottingham, Nottingham NG7 2RD, UK. Hauptthemen laut MEDLINE: Deep Learning, Automobile Driving. Die Arbeit wurde bisher 24-mal zitiert (Crossref) und nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: UK Research and Innovation, Horizon Centre for Doctoral Training at the University of Nottingham, Microlise.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

There is great interest in automatically detecting road weather and understanding its impacts on the overall safety of the transport network. This can, for example, support road condition-based maintenance or even serve as detection systems that assist safe driving during adverse climate conditions. In computer vision, previous work has demonstrated the effectiveness of deep learning in predicting weather conditions from outdoor images. However, training deep learning models to accurately predict weather conditions using real-world road-facing images is difficult due to: (1) the simultaneous occurrence of multiple weather conditions; (2) imbalanced occurrence of weather conditions throughout the year; and (3) road idiosyncrasies, such as road layouts, illumination, and road objects, etc. In this paper, we explore the use of a focal loss function to force the learning process to focus on weather instances that are hard to learn with the objective of helping address data imbalances. In addition, we explore the attention mechanism for pixel-based dynamic weight adjustment to handle road idiosyncrasies using state-of-the-art vision transformer models. Experiments with a novel multi-label road weather dataset show that focal loss significantly increases the accuracy of computer vision approaches for imbalanced weather conditions. Furthermore, vision transformers outperform current state-of-the-art convolutional neural networks in predicting weather conditions with a validation accuracy of 92% and an F1-score of 81.22%, which is impressive considering the imbalanced nature of the dataset.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningAutomobile DrivingAccidents, TrafficNeural Networks, ComputerWeather

Stichwörter der Autor:innen

autonomous vehiclescomputer visiondeep learningimage classificationloss functionsvision transformersweather detection

Autor:innen und Institutionen

  • School of Computer Science, University of Nottingham, Nottingham NG7 2RD, UK.

ORCID-Profile: Jimiama Mosima Mafeni Mase, Grazziela Figueredo

Förderung

  • UK Research and Innovation
  • Horizon Centre for Doctoral Training at the University of Nottingham
  • Microlise

Zitieren

APA

Samo, M., Mafeni Mase, J. M., & Figueredo, G. (2023). Deep Learning with Attention Mechanisms for Road Weather Detection. Sensors. https://doi.org/10.3390/s23020798

Vancouver

Samo M, Mafeni Mase JM, Figueredo G. Deep Learning with Attention Mechanisms for Road Weather Detection. Sensors. 2023. doi:10.3390/s23020798

BibTeX
@article{Samo2023,
  author = {Samo, Madiha and Mafeni Mase, Jimiama Mosima and Figueredo, Grazziela},
  title = {Deep Learning with Attention Mechanisms for Road Weather Detection},
  journal = {Sensors},
  year = {2023},
  doi = {10.3390/s23020798}
}
Kostenlos starten →