Deep Learning Approaches For Traffic Sign Detection

Authors

  • S Ramana Reddy Assitant Professor, Department of Artificial Intelligence and Data Science, Vignan Institute of Technology and Science, Hyderabad, India Author
  • Y Raj Kumar 2 UG Student, Department of AI&DS, Vignan Institute of Technology and Science, Hyderabad, India Author
  • T Sai Kiran UG Student, Department of AI&DS, Vignan Institute of Technology and Science, Hyderabad, India Author
  • T Tarun 2 UG Student, Department of AI&DS, Vignan Institute of Technology and Science, Hyderabad, India Author

DOI:

https://doi.org/10.33425/3066-1226.1134

Keywords:

Recognition of traffic signs, detection models, adaptive learning, pre-trained model, detection models

Abstract

Detecting traffic signs is important for many purposes, like ensuring safe driving and identifying illegal driving behavior. This study assesses specialized tools (e.g., Resnet, Inception, Mobile-net, and Dark net) and sophisticated detection models. It improves the trained prior model using the Microsoft Coco dataset for German car sign detection by using adaptive learning. evaluation of image size, computational complexity, speed, memory utilization, and accuracy. The results show that R-Fcn Resnet 101 balances precision and quick performance While Ssd Mobile-net is the fastest and best memory, Yolo Version 2 excels in performance & precision making it perfect for embedded devices, mobile devices, and gadgets.

Downloads

Published

2025-07-28

Issue

Vol. 5 No. 4

Section

Articles