Research Explorer

Project

Robust language diarization in reverberant and noisy environments
Code
01W06619
Duration
01 October 2019 → 01 October 2019
Funding
Regional and community funding: Special Research Fund
Research disciplines
  • Engineering and technology
    • Audio and speech computing
Keywords
Language diarization Spoken language identification Multichannel speech enhancement Multilingual automatic speech recognition
 
Project description
 

Language diarization refers to language recognition and segmentation of code-switched speech. This is critical for multilingual speech recognition systems. To increase the robustness of the diarization system to background interference, multi-channel speech enhancement techniques will be employed. The project aims to (a) develop language diarization system for Indian languages and (b) make it robust by multichannel speech enhancement techniques.