UDE diatoms in the wild 2024: a new image dataset of freshwater diatoms for training deep learning models
Largest available diatom image dataset from our ecological research projects
Highlights: 318 diatom samples, 83570 images if 611 diatom taxa, 101 species with at least 100 examples, two deep learning experiments going beyond standard supervised learning
Collaboration of the phycology lab of the university duisburg essen, the biodata mining group of Bielefeld university and the georgia tech group of the university of lorraine. Funded partially by the crc RESIST 1439 and the Humboldt foundation.
📢New paper out! The biggest diatom training set has been published on GigaScience. Read all about it here: doi.org/10.1093/giga...