Skip to Main content Skip to Navigation
Conference papers

Combinaison de mots et de syllabes pour transcrire la parole

Luiza Orosanu 1 Denis Jouvet 1
1 PAROLE - Analysis, perception and recognition of speech
Inria Nancy - Grand Est, LORIA - NLPKD - Department of Natural Language Processing & Knowledge Discovery
Abstract : Combining words and syllables for speech transcription This paper analyzes the use of hybrid language models for automatic speech transcription. The goal is to later use such an approach as a support for helping communication with deaf people, and to run it on an embedded decoder on a portable device, which introduces constraints on the model size. The main linguistic units considered for this task are the words and the syllables. Various lexicon sizes are studied by setting thresholds on the word occurrence frequencies in the training data, the less frequent words being therefore syllabified. Using this kind of language model, the recognizer can output between 69% and 96% of the words (whereas the other words, will be represented by syllables). By setting different thresholds on the confidence measures associated to the recognized words, the most reliable word hypotheses can be identified, and they have correct recognition rates between 70% and 92%.
Complete list of metadatas

Cited literature [20 references]  Display  Hide  Download

https://hal.inria.fr/hal-01080351
Contributor : Denis Jouvet <>
Submitted on : Wednesday, November 5, 2014 - 9:40:36 AM
Last modification on : Tuesday, December 18, 2018 - 4:38:02 PM
Document(s) archivé(s) le : Friday, February 6, 2015 - 10:11:23 AM

File

Luiza -- articleJEP2014 - envo...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01080351, version 1

Collections

Citation

Luiza Orosanu, Denis Jouvet. Combinaison de mots et de syllabes pour transcrire la parole. XXXème édition des Journées d'Etudes sur la Parole, Jun 2014, Le Mans, France. ⟨hal-01080351⟩

Share

Metrics

Record views

346

Files downloads

289