Vers la reconnaissance de mini-messages manuscrits

TL;DR

This paper investigates handwritten SMS recognition, focusing on unique language features like abbreviations and consonantal writing, and evaluates strategies to improve recognition accuracy using specialized language models.

Contribution

It introduces a study of handwritten SMS recognition emphasizing phenomena like consonant skeletons, rebus, and phonetic writing, with comparative analysis of recognition methods.

Findings

Standard recognition systems perform poorly on SMS-specific phenomena.

Using specialized language models improves recognition accuracy.

The study highlights the importance of tailored models for SMS handwriting.

Abstract

Handwriting is an alternative method for entering texts which composed Short Message Services. However, a whole new language features the texts which are produced. They include for instance abbreviations and other consonantal writing which sprung up for time saving and fashion. We have collected and processed a significant number of such handwritten SMS, and used various strategies to tackle this challenging area of handwriting recognition. We proposed to study more specifically three different phenomena: consonant skeleton, rebus, and phonetic writing. For each of them, we compare the rough results produced by a standard recognition system with those obtained when using a specific language model to take care of them.