Synthetic MSI Images of Georgian Palimpsests (SGP Dataset)

This is a dataset of Synthetic MSI Images of Georgian Palimpsests (SGP Dataset). It has been created for the purpose of training inpainting models in order to remove the overtext and reconstruct the undertext. It consists of three subsets: train, test and validation. Each synthetic palimpsest has its on mask and ground truth imagesas follows: Ground Truth Image: ImageName_a Mask Image: ImageName_b Synthetic Palimpsests: ImageName_c The typeface used to generate this synthetic dataset is for a very particular and unknown script to generate synthetic training samples. The first draft of the typeface was created by Jost Gippert in 2005, while the final version was prepared by Andreas Stötzner in 2007. The research for this work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany's Excellence Strategy – EXC 2176 ‘Understanding Written Artefacts: Material, Interaction and Transmission in Manuscript Cultures', project no. 390893796. The research was conducted within the scope of the Centre for the Study of Manuscript Cultures (CSMC) at Universität Hamburg.