Robust speech recognition

  title={Robust speech recognition},
  author={Richard M. Stern},
An Automatic Speech Recognition (ASR) component can degrade its onfield performance significantly if it was not designed taking into account the variabilities that can be found in the real-world. According to the early literature in the field, the term “robustness” was initially associated to the main variability factors in the input speech due to the environmental noise, which now is called as “environmental robustness”. Beside this, today other types of robustness are investigated as, for… CONTINUE READING