VTT-Coffee: A small dataset for instance segmentation and iterative learning

Tuomas Sormunen, Matti Okkonen, Arttu Lämsä, Miguel Bordallo LópezZenodo 2021 · DOI · OpenAlex

Abstract

VTT-Coffee is a small scale dataset containing microscopic images of coffee grinds and annotations of their contours (segmentation masks). The dataset is useful to benchmark instance segmentations methods, especially for use cases containing just one class with a lot of variability.
BibTeX
@inproceedings{sormunen2021vtt,
  author = {Tuomas Sormunen and Matti Okkonen and Arttu Lämsä and Miguel Bordallo López},
  title = {{VTT-Coffee: A small dataset for instance segmentation and iterative learning}},
  booktitle = {Zenodo},
  year = {2021},
  doi = {10.5281/zenodo.4980572},
}

Loading related content...

Figure