HyperKvasir: The Largest Gastrointestinal Dataset (Images) is a dataset for instance segmentation, semantic segmentation, object detection, classification, and semi supervised learning tasks. It is used in the medical industry.

The dataset consists of 111079 images with 2134 labeled objects belonging to 2 different classes including polyp and polyp_bbox.

Images in the HyperKvasir Images dataset have pixel-level instance segmentation and bounding box annotations. Due to the nature of the instance segmentation task, it can be automatically transformed into a semantic segmentation task (only one mask for every class). There are 110079 (99% of the total) unlabeled images (i.e. without annotations). There are 3 splits in the dataset: unlabeled-images (99417 images), labeled-images (10662 images), and segmented-images (1000 images). Alternatively, the dataset could be split into 2 GI tracts: lower-gi-tract (7210 images) and upper-gi-tract (3452 images), or into 4 categories of findings: anatomical-landmarks (4104 images), pathological-findings (2642 images), therapeutic-interventions (1991 images), and quality-of-mucosal-views (1925 images), or into 23 classification image sets: bbps-2-3 (1148 images), polyps (1028 images), cecum (1009 images), dyed-lifted-polyps (1002 images), pylorus (999 images), dyed-resection-margins (989 images), z-line (932 images), retroflex-stomach (764 images), bbps-0-1 (646 images), ulcerative-colitis-grade-2 (443 images), esophagitis-a (403 images), retroflex-rectum (391 images), esophagitis-b-d (260 images), ulcerative-colitis-grade-1 (201 images), ulcerative-colitis-grade-3 (133 images), impacted-stool (131 images), barretts-short-segment (53 images), barretts (41 images), ulcerative-colitis-grade-0-1 (35 images), ulcerative-colitis-grade-2-3 (28 images), ulcerative-colitis-grade-1-2 (11 images), ileum (9 images), and hemorrhoids (6 images). The dataset was released in 2020 by the NO-SW-AU joint research group.

Here are the visualized examples for the classes:

HyperKvasir: The Largest Gastrointestinal Dataset (Images) is under CC BY 4.0 license.

Source

If you make use of the HyperKvasir Segmentation data, please cite the following reference:

@article{Borgli2020,
  title = {{HyperKvasir, a comprehensive multi-class
    image and video dataset for gastrointestinal endoscopy}},
  author = {
    Borgli, Hanna and Thambawita, Vajira and
    Smedsrud, Pia H and Hicks, Steven and Jha, Debesh and
    Eskeland, Sigrun L and Randel, Kristin Ranheim and
    Pogorelov, Konstantin and Lux, Mathias and
    Nguyen, Duc Tien Dang and Johansen, Dag and
    Griwodz, Carsten and Stensland, H{\aa}kon K and
    Garcia-Ceja, Enrique and Schmidt, Peter T and
    Hammer, Hugo L and Riegler, Michael A and
    Halvorsen, P{\aa}l and de Lange, Thomas
  },
  doi = {10.1038/s41597-020-00622-y},
  issn = {2052-4463},
  journal = {Scientific Data},
  number = {1},
  pages = {283},
  url = {https://doi.org/10.1038/s41597-020-00622-y},
  volume = {7},
  year = {2020}
}

Source

Dataset HyperKvasir Images can be downloaded in Supervisely format:

As an alternative, it can be downloaded with dataset-tools package:

pip install --upgrade dataset-tools

… using following python code:

import dataset_tools as dtools

dtools.download(dataset='HyperKvasir Images', dst_dir='~/dataset-ninja/')

Make sure not to overlook the python code example available on the Supervisely Developer Portal. It will give you a clear idea of how to effortlessly work with the downloaded dataset.

The data in original format can be downloaded here.