Handling Heavily Abbreviated Manuscripts: HTR engines vs text normalisation approaches

Although abbreviations are fairly common in handwritten sources, particularly\nin medieval and modern Western manuscripts, previous research dealing with\ncomputational approaches to their expansion is scarce. Yet abbreviations\npresent particular challenges to computational approaches such as handwritten\ntext recognition and natural language processing tasks. Often, pre-processing\nultimately aims to lead from a digitised image of the source to a normalised\ntext, which includes expansion of the abbreviations. We explore different\nsetups to obtain such a normalised text, either directly, by training HTR\nengines on normalised (i.e., expanded, disabbreviated) text, or by decomposing\nthe process into discrete steps, each making use of specialist models for\nrecognition, word segmentation and normalisation. The case studies considered\nhere are drawn from the medieval Latin tradition.\n