The Declaration on Research Assessment (DORA) developed in 2012 suggested recognizing additional datasets as important research contributions besides peer-review articles. Providing open access to research data is now a requirement in, for instance, projects funded by the Research Council of Norway or EU.
UiT and CSG comply with and endorse the principle of providing open access to data resulting from our research. This is often referred to as the FAIR principle, and we contribute to this mandatory requirement managing our research data (and affilited code) through the UiT Research Data Portal.
Additionally, we are collaborating with our inter-disciplinary colleagues in curating, specifically preparing, and publishing annotated datasets for others to use. Particularly, development of modern analytics software like machine-learning (ML) and Deep Learning ML depends on reliable, broad datasets for training and evaluation purposes. We have already published labeled datasets in the gastrointestinal medicine and sports science domains to contribute to scientific advancement in the fields.
- “Soccer video and player position dataset” (ACM, 2014)
- “KVASIR: A Multi-Class Image Dataset for Computer Aided Gastrointestinal Disease Detection” (ACM, 2017)
- “Kvasir SEG: A segmented polyp dataset” (Springer, 2020)
- “HyperKvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy” (Nature Scientific Data, 2020)
- “PMData: a sports logging dataset” (ACM, 2020)
- “Kvasir-Instrument: Diagnostic and therapeutic tool segmentation dataset in gastrointestinal endoscopy” (OSF, Springer 2021)
-
“Kvasir-Capsule, a video capsule endoscopy dataset” (Nature Scientific Data, 2021)