Automatic speech recognition

José Adrián Rodríguez Fonollosa



In speech recognition, the following lines of research are being undertaken:

  • Large vocabulary systems
  • Generation of confidence measures
  • Multi-dialect and multilingual recognition systems
  • Robust systems that combat noise and environmental variations
  • Integration in multimedia environments
  • Recognition in rooms of signals from clusters of microphones
  • Language Modeling
  • Integration in speech translation systems

 

One of the ultimate goals of this research is to improve the performance of large vocabulary automatic speech recognition systems to obtain high quality speech-to-speech multilingual translation systems. A specific application would be the translation of parliamentary speeches.

Another goal is to attain good performance in the automatic recognition of speech signals received in a room with a number of microphones. This line of research has the additional advantage of being able to detect other kinds of noises (background noise, music, applause, etc.). In order to achieve better performance, particularly when a number of noises and voice signals overlap, a blind separation of sources must be undertaken. This involves processing microphone clusters and selecting a robust set of signal parameters.