The aim of this compute project is to develop and evaluate models for the recognition and processing of historical writings. Besides handwritten text recognition,layout analysis and segmentation, and document image preprocessing will be investigated, with a special focus on low resource datasets, such as historical ciphers.