Automatic Abbreviation Detection in Medieval Medical Documents

Abstract

Automatic palaeographic transcription of medieval manuscripts is a challenging task. Transcribing and deciphering these manuscripts is considered complicated and problematic, especially due to particular characteristics of medieval handwriting and the composition and conservation of these documents. In this paper, a new problem in handwriting recognition is addressed: the detection of abbreviations in medieval medical manuscripts. We propose a histogram-based method to detect these specific kinds of characters. The results achieved with real medieval medical manuscripts are very promising, and it opens new research lines in handwriting recognition.