Natural Language Inspired Approach for Handwritten Text Line Detection in Legacy Documents∗

Document layout analysis is an important task needed for handwritten text recogni-tion among other applications. Text lay-out commonly found in handwritten legacy documents is in the form of one or more paragraphs composed of parallel text lines. An approach for handwritten text line de-tection is presented which uses machine-learning techniques and methods widely used in natural language processing. It is shown that text line detection can be accu-rately solved using a formal methodology, as opposed to most of the proposed heuris-tic approaches found in the literature. Ex-perimental results show the impact of us-ing increasingly constrained ”vertical lay-out language models ” in text line detection accuracy.