CART: CARLA Town 10 Dataset

Abhijith Sharma*,1, Apurva Narayan†,2, Nasser Azad†,1, Sebastian Fischmeister†,1, Stefan Marksteiner3
1University of Waterloo, 2Western University, 3AVL, Graz Austria
*Indicates Main Author, †Indicates Equal Advising
Dataset (zip) Dataset (tar) Original Paper
Inference Overview

Sample image frames from the CART (CARLA Town 10) dataset. Town 10 is the HD map in CARLA, replicating realistic scenarios.

Data Description

  • Photo-realistic dataset out of CARLA's Town 10
  • Total of 4500 images (Train ~ 4300, Val ~ 200) with 1280×720 pixels resolution.
  • Classes include: Person, Vehicle, Motorbike, Traffic Light, Stop Sign.
  • Collected in CARLA using autopilot driving in Town 10 (moving camera scenario).
  • Images are collected under diverse lighting (noon, sunset, night) and weather conditions (rain, clear, fog)
  • The traffic level is moderate with around 40 vehicles and 70 pedestrians in the map.
  • The labels are available in YOLO format and annotations are performed using the CVAT tool: https://www.cvat.ai/

This dataset has been used in original paper to train YOLOv5 model as well as the adversarial patch used in the experiments

BibTeX

If you find our dataset useful, please consider citing our work.

@article{sharma2024avatar,
      title={AVATAR: Autonomous Vehicle Assessment through Testing of Adversarial Patches in Real-time},
      author={Sharma, Abhijith and Narayan, Apurva and Azad, Nasser Lashgarian and Fischmeister, Sebastian and Marksteiner, Stefan},
      journal={IEEE Transactions on Intelligent Vehicles},
      year={2024},
      publisher={IEEE}
    }

Acknowledgements

We extend our gratitude to the members of the Real-time Embedded Software Group @UWaterloo for their invaluable insights, critical brainstorming sessions, and innovative ideas that greatly contributed to this work.