Using the MGGI Methodology for Category-Based Language Modeling in Handwritten Marriage Licenses Books

Abstract

IEEERomero Gómez, V.; Fornes, A.; Vidal Ruiz, E.; Sánchez Peiró, JA. (2016) Abstract-Handwritten marriage licenses books have been used for centuries by ecclesiastical and secular institutions to register marriages. The information contained in these historical documents is useful for demography studies and genealogical research, among others. Despite the generally simple structure of the text in these documents, automatic transcription and semantic information extraction is difficult due to the distinct and evolutionary vocabulary, which is composed mainly of proper names that change along the time. In previous works we studied the use of category-based language models to both improve the automatic transcription accuracy and make easier the extraction of semantic information. Here we analyze the main causes of the semantic errors observed in previous results and apply a Grammatical Inference technique known as MGGI to improve the semantic accuracy of the language model obtained. Using this language model, full handwritten text recognition experiments have been carried out, with results supporting the interest of the proposed approach.

Joint Recognition of Handwritten Text and Named Entities with a Neural End-to-End Model

“…It is definitely difficult to transcribe certain names that have never been seen before. The proposed approach could be combined with a category-based language model [1] which could potentially improve the results. Our best performing model took 4 hours 38 to run 133 training epochs with a NVIDIA GTX 1080 GPU.…”

“…Extracting information from historical handwritten text documents in an optimal and efficient way is still a challenge to solve, since text in these kind of documents are not as simple to read as printed characters or modern handwritten calligraphies [1], [2]. Historical manuscripts contain information that gives an interpretation of the past of societies.…”

“…Many existing solutions make use of Artificial Neural Networks (ANNs) to transcribe handwritten text lines and then parse the transcribed text with a Named Entity Recognition model, but the precision of those existing solutions is still to improve [1], [2], [4]. One possible approach is to start with already segmented words, by an automatic or manual process, and predict the semantic category using visual descriptors (c.f.…”