Automatic Speech Recognition For Polish In A Computer Game Interface

Artur Janicki , Dariusz Wawer

Abstract

The paper describes the process of designing a task-oriented continuous speech recognition system for Polish, based on CMU Sphinx 4, to be used in the voice interface of a computer game called Rally Navigator. The concept of the game is presented, the stages of creating the acoustic model and the language model are described in details, taking into account the specificity of the Polish language. Results of initial experiments show that as little as 15 minutes of audio material is enough to produce a highly effective single-speaker command-and-control ASR system for the computer game, providing the sentence recognition accuracy of 97.6%. Results of the system adaptation for a new speaker are presented. It is also showed that the statistic trigram-based language model with negative trigrams yields the best recognition results.
Author Artur Janicki IT
Artur Janicki,,
- The Institute of Telecommunications
, Dariusz Wawer IT
Dariusz Wawer,,
- The Institute of Telecommunications
Pages711 - 716
Book Ganzha Maria, Maciaszek Leszek A., Paprzycki Marcin: Proceedings of the 2011 Federated Conference on Computer Science and Information Systems, 2011, IEEE, ISBN 978-1-4577-0041-5, [978-83-60810-35-4], 1115 p.
URL Automatic speech recognition for polish in a computer game interface
Languageen angielski
File
06078265.pdf (file archived - login or check accessibility on faculty) 06078265.pdf 163.99 KB
Score (nominal)10
Citation count*4 (2018-02-16)
Cite
Share Share



* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.
Back