Semantic segmentation of epithelial tissues in digitalised histopathological images streamlines cancer diagnosis and prognosis by enhancing accuracy and efficiency, reducing the manual effort of professional pathologists and speeding up the analysis process. Brázdil et al. [1] presented a re-staining-based method for automatic epithelium annotation to train an AI model. The thesis aims to design, train, and evaluate several U-net-based neural networks that segment epithelial tissue of histopathological colorectal and breast tissue images provided by Masaryk Memorial Cancer Institute. The student will familiarise themself and present the basics of digital pathology, semantic segmentation and the methods of [1]. They preprocess the data, train, and evaluate several variants of segmentation models in the cloud infrastructure of the RationAI group. The student will optionally contribute the code to HistoPipe, the RationAI open-source library for digital histopathology. [1] Brázdil et al. Automated annotations of epithelial cells and stroma in hematoxylin–eosin-stained whole-slide images using cytokeratin re-staining. The Journal of Pathology: Clinical Research. 2022.