WEDGE: A multi-weather autonomous driving dataset built from generative vision-language models

Aboli Marathe1, Deva Ramanan1, Rahee Walambe2, Ketan Kotecha2,
1Carnegie Mellon University 2 Symbiosis Centre for Applied AI (SCAAI), Symbiosis International University (SIU), India, Symbiosis Institute of Technology (SIT), Symbiosis International University (SIU), India
Paper arXiv Code Data
Image Grid with Captions
Image 1
WEDGE Dataset Samples
Image 2
Sim2Real inference for Object Detection
Image 3
Sim2Real Image Similarity Distributions
Image 4
Anomalies and Interesting Examples of WEDGE

BibTeX

@misc{marathe2023wedge,
     @InProceedings{Marathe_2023_CVPR,
    author    = {Marathe, Aboli and Ramanan, Deva and Walambe, Rahee and Kotecha, Ketan},
    title     = {WEDGE: A Multi-Weather Autonomous Driving Dataset Built From Generative Vision-Language Models},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
    month     = {June},
    year      = {2023},
    pages     = {3317-3326}

}