Omniglot dataset

doi:doi:10.57702/jp2k7lbz

Omniglot dataset

The Omniglot dataset consists of 100 classes, each containing 20 images. Ten images were taken from each class for augmentation, and the rest were used as the test set. Each image can generate 2004 new images by the ND-MLS method, so 2 million images were used for the training model.

BibTex:

@dataset{Arman_Kazemi_and_Mohammad_Mehdi_Shari_and_Ann_Franchesca_Laguna_and_Franz_M¨uller_and_Ramin_Rajaei_and_Ricardo_Olivo_and_Thomas_K¨ampfe_and_Michael_Niemier_and_X_Sharon_Hu_2024,
    abstract = {The Omniglot dataset consists of 100 classes, each containing 20 images. Ten images were taken from each class for augmentation, and the rest were used as the test set. Each image can generate 2004 new images by the ND-MLS method, so 2 million images were used for the training model.},
    author = {Arman Kazemi and Mohammad Mehdi Shari and Ann Franchesca Laguna and Franz M¨uller and Ramin Rajaei and Ricardo Olivo and Thomas K¨ampfe and Michael Niemier and X. Sharon Hu},
    doi = {10.57702/jp2k7lbz},
    institution = {No Organization},
    keyword = {'Computer Vision', 'Handwritten Characters', 'Omniglot dataset', 'alphabets', 'character recognition', 'cluster evaluation', 'data augmentation', 'handwritten characters', 'image clustering'},
    month = {dec},
    publisher = {TIB},
    title = {Omniglot dataset},
    url = {https://service.tib.eu/ldmservice/dataset/omniglot-dataset},
    year = {2024}
}