1. Introduction: “Rescapé”
The Rescapé project is a collaboration at PSL University between the Centre Jean-Mabillon at the École nationale des chartes and EquipEx Biblissima+ at Campus Condorcet. Funded by Biblissima+, Rescapé sought to apply HTR to the fire-damaged T manuscript. A machine-learning approach for segmenting and transcribing images of handwritten documents, the purpose of Rescapé is to avail of the affordances of HTR technology to recognise both the layout structures and the textual content of this fire-damaged medieval manuscript. The segmentation process is the first stage in the HTR approach and focuses on analysing the layout of a manuscript, in other words, recognising the different regions on a manuscript page. Once the segmentation stage is completed satisfactorily, the second and final stage is the transcription process which consists of identifying and transcribing the textual content of the manuscript. While the manuscript’s sizeable corpus justified the use ...


