San Vitale challenge: Automatic Reconstruction of Ancient Colored Glass Windows

The sixth-century Basilica of San Vitale in Ravenna, Italy, once featured intricate circular colored glass windows that illuminated its interior. Although these windows are now lost, several fragments were recovered during recent restorations.

Unfortunately, reconstructing the original glass windows from these fragments is extremely complex and time-consuming, requiring the use of specialized expertise. Therefore, the development of automatic reconstruction techniques based on Artificial Intelligence is particularly important and challenging, due to, for instance, the presence of uniform color, damaged glass edges, and many fragment outliers.


In this direction, the San Vitale Challenge was organized to gather the best methods and algorithms. The dataset collected for the challenge is available for all researchers to contribute to this field.

The link to the website of the challenge is the following: https://sites.google.com/view/svitale-2024

Dataset

    • Many pictures of 15 disks have been collected 
      • 10 for training
      • 5 for test
      • 5 distinct colors
      • Each fragment is individually identified and photographed
  • Controlled acquisition environment
    • Fixed distance from table
    • Uniform LED lighting
    • We include color checker and ruler
    • Minimal processing on RAW images
    • Barrel distortion correction
    • Color balancing
  • Human annotations
    • We provide manually annotated polylines for each fragment
    • Synthetic description of the fragment’s border
    • Encoded as a list of vertices in XY coordinates
    • Starting from the top in clockwise order
    • Adjacency lists for disks of training set
    • Indicates which fragments are adjacent to which other fragments
    • Goal is to output an adjacency list for the 5 disks in the test set

 

Download

The dataset will be released soon!

License and Reference

The SVitale dataset is released under the CC BY-NC 4.0 license: you may not use the material for commercial purposes and you must give appropriate credit (see the reference below). You are free to download and modify the dataset.

If you use this dataset or you want to see the challenge results, please read and cite the following paper:

@inproceedings{di2024svitale,
  title={San Vitale Challenge: Automatic Reconstruction of Ancient Colored Glass Windows},
  author={Di Domenico, Nicol{\`o} and Borghi, Guido and Franco, Annalisa and Boschetti, Marco et al.},
  booktitle={The 18th European Conference on Computer Vision Wrokshop 2024},
  year={2024}
}