Marble Crack Segmentation (MCS) Dataset

Marble is a metamorphic, i.e., heated and compressed, rock originating from sedi-mentary limestone. Marble is quarried naturally and contains mainly calcium. In some cases, magnesium can prevail, and then the rock is called dolomite. Marble is quarried in blocks of 3×2×5 meters. From these blocks, slabs are cut, usually in the size of 3×2 me-ters and thickness ranging from 2-4 cm. Slabs are further cut into tiles in various sizes, depending on the needs of the construction.

The detection of flaws using computer vision is a difficult task due to the complex textures on many marble types marketed today. It is very common to encounter Type I cracks that were filled naturally with minerals. These types of filled cracks are called fis-sures and don’t create problems on the stability of the slab (as seen in the figure below). Due to their optical similarity to cracks, fissures pose a significant problem to the automatic recognition of cracks.

Dataset Properties

Total samples: 246
Number of classes: 2 (0: background, 1: crack)
Image format: jpg/png
Image size: 256xx256

Structure

.
├── dataset
    ├── images
      ├── _0_0_20210531_17292.jpg
      ├── _0_0_20210531_17293.jpg
      ...
    ├── masks
      ├── _0_0_20210531_17292.png
      ├── _0_0_20210531_17293.png
      ...

Each image's corresponding mask has the same name (but with a png extention).

Acknowledgements

We built upon the original dataset, called Marble Surface Anomaly Detection - 2 and uploaded by Aman Rastogi to kaggle, by annotating the marbles with cracks for semantic segmenation.

Citation

If you use the dataset in a scientific publication, please cite us using the following bibtex citation:

@Article{electronics11203289,
    AUTHOR = {Vrochidou, Eleni and Sidiropoulos, George K. and Ouzounis, Athanasios G. and Lampoglou, Anastasia and Tsimperidis, Ioannis and Papakostas, George A. and Sarafis, Ilias T. and Kalpakis, Vassilis and Stamkos, Andreas},
    TITLE = {Towards Robotic Marble Resin Application: Crack Detection on Marble Using Deep Learning},
    JOURNAL = {Electronics},
    VOLUME = {11},
    YEAR = {2022},
    NUMBER = {20},
    ARTICLE-NUMBER = {3289},
    URL = {https://www.mdpi.com/2079-9292/11/20/3289},
    ISSN = {2079-9292},
    DOI = {10.3390/electronics11203289}
}

Also cite the original dataset, using the following bibtex:

@misc{rastogi-a,
  author = {Rastogi, A.kaggle},
  title = {Marble Surface Anomaly Detection - 2 Available online},
  url = {https://www.kaggle.com/datasets/wardaddy24/marble-surface-anomaly-detection-2},
  language = {en}
}

License

Attribution 4.0 International