Dataset used in training for parts of OntoUML Image Taxonomy Extractor (OITE)
Half of this dataset is composed of the 643 OntoUML images from the OntoUML Catalogue, and the other half comprised 321 randomly selected class diagrams and 322 randomly selected non-UML images from the dataset of the paper “Multiclass classification of four types of UML diagrams from images using deep learning". The idea behind it was to have equal parts OntoUML and non-OntoUML images. In the paper it is referred to as the “binary dataset”.
See: https://github.com/SimeonKaishev/OntoUML_IMG_Converter