Robust parameterizations of Automatic Speech Recognition (ASR) in communications networks

In this paper we address the problem of Automatic Speech Recognition (ASR) when the speech signal has been transmitted over communications networks. In these conditions, the main causes of distortion in an ASR system are ambient noise, transmission errors and the encoding-decoding process [32]. In the literature we are able to find multiple solutions for this problem, from different points of views; however, in this paper we will focus the analysis on solutions with robust parameterizations for the above distortions.

• Tipo de recurso: Article of journal
• Fecha de publicación: 2014
• Institución: Universidad Antonio Nariño
• Repositorio: Repositorio UAN
• Idioma: spa
• OAI Identifier: oai:repositorio.uan.edu.co:123456789/10433
• Acceso en línea: https://revistas.uan.edu.co/index.php/ingeuan/article/view/356; https://repositorio.uan.edu.co/handle/123456789/10433
• Palabra clave: ASR; Speech Coding; CELP coders; packet networks; VolP; transmission errors; packet loss; noise; mobile networks; UMTS; LTE
• License: https://creativecommons.org/licenses/by-nc-sa/4.0