The repository contains the Kvasircapsule-seg dataset. It includes 55 medically verified video capsule frames polyp frames from Kvasir-Capsule dataset. We have annotated the polyp class of KvasirCapsule and generated corresponding ground truth masks. This dataset can be used for semantic meta-learning, domain generalization, semantic segmentation, and out-of-distribution detection task.
The official website for the dataset is: https://datasets.simula.no/kvasir-capsule-seg/
The official publication link: https://ieeexplore.ieee.org/document/9474743
The Kvasir-Capsule can be downloaded this github repository and the dropbox link: https://www.dropbox.com/home/KvasirCapsule-SEG.
It is used for capturing video capsule frames. The examples of capsule endocopy polyp frames are below:Please cite our paper if you find the work useful:
@proceedings{jha2021nanonet, title={NanoNet: Real-Time Polyp Segmentation in Endoscopy}, author={Jha, Debesh and Tomar, Nikhil Kumar and Ali, Sharib and Riegler, Michael A and Johansen, H{\aa}vard D and Johansen, Dag and Halvorsen, P{\aa}l}, booktitle={Proceedings of the IEEE Computer Based Multimedia System}, year={2021}, publisher={IEEE} }
It is an open-access dataset which can be downloaded for research and academic purpose. For industrial purpose, a prior consent is required before using them.
Please contact debesh.jha@northwestern.edu for any further questions.