SkyFinder Dataset

This dataset consists of images from 53 AMOS webcams (https://mvrl.github.io/AMOS). The image sequences were selected so that they were stable so that a single sky/ground mask could be used for all images in a given camera.

Contents

• [integer camera ID].zip : contains images from each camera
• skyfinder_masks.zip : contains the sky/ground masks

Additional Links

• Project Webpage: https://cs.valdosta.edu/~rpmihail/skyfinder/
• Another Dataset Summary: https://mvrl.github.io/SkyFinder

Citation

If you use this dataset, please cite the following paper:

Mihail RP, Workman S, Bessinger Z, Jacobs N. 2016. Sky Segmentation in the Wild: An Empirical Study. In: IEEE Winter Conference on Applications of Computer Vision (WACV). 1–6. DOI: 10.1109/WACV.2016.7477637.

@inproceedings{mihail2016sky,
  author = {Mihail, Radu Paul and Workman, Scott and Bessinger, Zach and Jacobs, Nathan},
  booktitle = {IEEE Winter Conference on Applications of Computer Vision (WACV)},
  pages = {1--6},
  doi = {10.1109/WACV.2016.7477637},
  title = {Sky Segmentation in the Wild: An Empirical Study},
 year = {2016}
}