Automating live and batch subtitling of multimedia contents for several European languages

Authors: Aitor Álvarez Muniain Carlos Mendes Metteo Raffaelli Thiago Luís Sérgio Paulo Nicola Piccinini Haritz Arzelus Irazusta Joao Neto Carlo Aliprandi Arantza del Pozo Echezarreta

Date: 01.09.2016


PDF

Abstract

The subtitling demand of multimedia content has grown quickly over the last years, especially after the adoption of the new European audiovisual legislation, which forces to make multimedia content accessible to all. As a result, TV channels have been moved to produce subtitles for a high percentage of their broadcast content. Consequently,the market has been seeking subtitling alternatives more productive than the traditional manual process. The large effort dedicated by the research community to the development of Large Vocabulary Continuous Speech Recognition (LVCSR) over the last decade has resulted in significant improvements on multimedia transcription, becoming the most powerful technology for automatic intralingual subtitling. This article contains a detailed description of the live and batch automatic subtitling applications developed by the SAVAS consortium for several European languages based on proprietary LVCSR technology specifically tailored to the subtitling needs, together with results of their quality evaluation.

BIB_text

@Article {
title = {Automating live and batch subtitling of multimedia contents for several European languages},
pages = {10823-108},
number = {18},
volume = {75},
keywds = {

Multimedia communication; Multimedia systems; Automatic speech recognition; Automatic subtitling; Subtitling quality; Access services


}
abstract = {

The subtitling demand of multimedia content has grown quickly over the last years, especially after the adoption of the new European audiovisual legislation, which forces to make multimedia content accessible to all. As a result, TV channels have been moved to produce subtitles for a high percentage of their broadcast content. Consequently,the market has been seeking subtitling alternatives more productive than the traditional manual process. The large effort dedicated by the research community to the development of Large Vocabulary Continuous Speech Recognition (LVCSR) over the last decade has resulted in significant improvements on multimedia transcription, becoming the most powerful technology for automatic intralingual subtitling. This article contains a detailed description of the live and batch automatic subtitling applications developed by the SAVAS consortium for several European languages based on proprietary LVCSR technology specifically tailored to the subtitling needs, together with results of their quality evaluation.


}
isi = {1},
doi = {10.1007/s11042-015-2794-z},
date = {2016-09-01},
year = {2016},
}
Vicomtech

Parque Científico y Tecnológico de Gipuzkoa,
Paseo Mikeletegi 57,
20009 Donostia / San Sebastián (Spain)

+(34) 943 309 230

close overlay