Books of Hours. the First Liturgical Data Set for Text Segmentation.

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 776–784 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 776 Books of Hours: the First Liturgical Corpus for Text Segmentation Amir Hazem*, B´eatrice Daille*, Marie-Laurence Bonhomme†, Martin Maarand†, M´elodie Boillet†‡, Christopher Kermorvant†‡ and Dominique Stutzmann§ * LS2N - Universit´e de Nantes, Nantes § Institut de recherche et d’histoire des textes (IRHT), Paris † TEKLIA, Paris ‡ LITIS - Universit´e de Rouen-Normandie, Rouen amir.hazem@ls2n.fr, beatrice.daille@univ-nantes.fr, Dominique.Stutzmann@irht.cnrs.fr, Kermorvant@teklia.com Abstract The Book of Hours was the bestseller of the late Middle Ages and Renaissance. It is a historical invaluable treasure, documenting the devotional practices of Christians in the late Middle Ages. Up to now, its textual content has been scarcely studied because of its manuscript nature, its length and its complex content. At first glance, it looks too standardized. However, the study of book of hours raises important challenges: (i) in image analysis, its often lavish ornamentation (illegible painted initials, line-fillers, etc.), abbreviated words, multilingualism are difficult to address in Handwritten Text Recognition (HTR); (ii) its hierarchical entangled structure offers a new field of investigation for text segmentation; (iii) in digital humanities, its textual content gives opportunities for historical analysis. In this paper, we provide the first corpus of books of hours, which consists of Latin transcriptions of 300 books of hours generated by Handwritten Text Recognition (HTR) - that is like Optical Character Recognition (OCR) but for handwritten and not printed texts. We designed a structural scheme of the book of hours and annotated manually two books of hours according to this scheme. Lastly, we performed a systematic evaluation of the main state of the art text segmentation approaches. Keywords: text segmentation, books of hours, structural scheme, hierarchical segmentation 1. Introduction In the late medieval Europe, the book of hours was a Chris- tian devotional manuscript used by lay people as a guid- ance book in their daily prayers (Leroquais, 1927; Wieck et al., 1988). While this personal object of about 300 pages was mainly used by aristocracy and, later on, by the mid- dle class, the emergence of the printing press made books of hours available for nearly every European family (Hind- man and Marrow, 2013). With a massive production result- ing in more than 10,000 extant manuscripts, this number one liturgical best seller is an invaluable source of infor- mation of the late medieval usage and practices and wit- ness to the profound changes in the European society on cultural, religious, and industrial levels. Since they were illuminated with beautiful miniature paintings and deco- rations (De Hamel, 1992), books of hours have attracted great attention. They became very popular for art histo- rians and many studies have been conducted for aesthetic purposes (Wieck, 2001). However, their textual content is still scarcely studied, often because of their length, the complexity of their liturgical content and the fact that they appear as being too standardized. Despite the standardized character of books of hours in …

Content truncated to 500 words