Additional BDD100K Context Annotations
Date
2023-07-24Author
Heidecker, Florian
Contributing Person/Institution
ProjectLeader: Fuchs, Erich
ProjectLeader: Sick, Bernhard
ProjectMember: Susetzky, Tobias
Metadata
Show full item recordAbstract
This dataset provides additional context annotations and extends the BDD100K dataset [1]. Therefore, the approximately 80,000 images with annotation for 2D object detection in the BDD100K dataset were annotated with additional context attributes. The application possibilities of the dataset are diverse. It could be used for model training or evaluating the model performance in different contexts combination, as we did. The annotated contexts per image contain 11 attributes (time_of_day, sky, illumination, precipitation, infrastructure, road, tunnel, construction_site, clear_windshield, light_exposure, and reflections). The underlying annotation guideline with further details is also available for download. To use the context annotations, the official [BDD100K Dataset](https://doc.bdd100k.com/download.html) is required. On top, the provided python-script "extent_bdd_with_context.py" has to be executed to merge the BDD100K original annotations and our additional context annotations. Further instructions are available in the readme file. [1] Yu, F. and Chen, H. and Wang, X. and Xian, W. and Chen, Y. and Liu, F. and Madhavan, V. and Darrell, T.: "BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learning", in Proc. of CVPR, Seattle, WA, USA, 2020, doi: 10.1109/CVPR42600.2020.00271 IMPORTANT: In case you use the dataset, please cite our corresponding article mentioned below. Version 1.0
Funder
BMWK - Bundesministerium für Wirtschaft und KlimaschutzRelated Resources
IsCitedBy: (Keine Angabe) Heidecker, F. and Susetzky, T. and Fuchs, E. and Sick, B.: Context Information for Corner Case Detection in Highly Automated Driving, in Proc. of ITSC, 2023, (accepted)Collections
Except where otherwise noted, this item's license is described as Creative Commons Attribution-NonCommercial 4.0