Traffic Sign Detection under Challenging Conditions

Dogancan Temel, Min-Hung Chen, Tariq Alshawi, Ghassan AlRegib

Last updated on Jun 23, 2020

Motivation

Existing traffic sign datasets are limited in terms of type and severity of challenging conditions. Metadata corresponding to these conditions are unavailable and it is not possible to investigate the effect of a single factor because of simultaneous changes in numerous conditions. Therefore, we introduce the CURE-TSD dataset, including various challenging conditions with both real and synthetic data.

Dataset Overview

Video number: 5733 videos
Video length: 300 frames/video (~1.7M total frames in the dataset)
Challenge types: 12
Challenge levels: 5
Challenging conditions (horizontal: challenge levels. vertical: challenge types).
Traffic sign types: 14
Traffic signs (Row 1: real signs. Row 2: synthetic signs).

Demo Videos

Real data:
Synthetic data:

Please check our paper for more results.

Resources

Papers & GitHub

TITS'19	arXiv'19

Download

To download the dataset, please visit our GitHub or IEEE DataPort.

BibTex

@article{temel2019traffic,
  title={Traffic sign detection under challenging conditions: A deeper look into performance variations and spectral characteristics},
  author={Temel, Dogancan and Chen, Min-Hung and AlRegib, Ghassan},
  journal={IEEE Transactions on Intelligent Transportation Systems (TITS)},
  year={2019},
  publisher={IEEE}
}

@article{temel2019challenging,
  title={Challenging environments for traffic sign detection: Reliability assessment under inclement conditions},
  author={Temel, Dogancan and Alshawi, Tariq and Chen, Min-Hung and AlRegib, Ghassan},
  journal={arXiv preprint arXiv:1902.06857},
  year={2019},
  url={https://arxiv.org/abs/1902.06857}
}

Members

Georgia Institute of Technology

Dogancan Temel	Min-Hung Chen	Tariq Alshawi	Ghassan AlRegib

Traffic Sign Detection under Challenging Conditions

Motivation