I am searching an open data set and/or a model "ready" (or close to be ready) to segment the sky out of outdoor images, e.g.:


Pascal Debrunner Oeschinen Lake, Kandersteg, Switzerland

Source; unsplash, Pascal Debrunner, Oeschinen Lake, Kandersteg, Switzerland

Desired output:

Desired output


Do you know some data sets or models on that topic?
Or better, meta-surveys on that question?

Info; the deep-learning backend will normally be tensorflow/keras based.

  • I wonder the Title of the question and questions inside the body seem to be different..!!
    – Orophile
    Jul 6, 2020 at 9:09
  • Indeed, but instead of a suite of Answers on specific data sets (which would be fine), if someone passing here knows a meta survey on that topic, that would be even better.
    – s.k
    Jul 8, 2020 at 9:51

1 Answer 1


Sky Dataset

A Dataset for Sky Segmentation

The Sky dataset contains a collection of 60 images with ground truth for sky segmentation. It was based on the Caltech Airplanes Side dataset by R. Fergus 15/02/03. Those images from the dataset which contained sky regions were selected and ground truths were created for them. The original dataset image names were kept unchanged.

enter image description here

The dataset can be downloaded here.

A large dataset of webcam images annotated with sky regions.

Download Here

Useful Resource

  • 60K images you mean?
    – s.k
    Jul 8, 2020 at 9:54
  • No, it's just 60 images
    – Orophile
    Jul 8, 2020 at 10:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.