Alexandre Girardi

  • Citations Per Year
Learn More
In speaker independent speech recognition, one problem we often face is the insu cient database for training. A system trained only with several male and female databases will likely lack the information which is present in speakers with di erent pitch and vocal tract lengths. In an extreme case, children database, that is not easy to obtain, is a good(More)
At the heart of the European Rail Train Management System (ERTMS) is the European Train Control System (ETCS). One major goal of the ERTMS-ETCS project is the standardization and unification of all train control and command systems in Europe. Hence, it is critical to have a reliable test bed for ease of validation and certification, enforcing the(More)
Man-Machine Interaction using only speech input is not well received by users, even for high performance recognizers (WER of about 2%). In most free text dictation application, attaining users intention is more important than specific speech tools performance, and low transaction success rate results in user’s rejection to speech interfaces [6]. For(More)
This paper describes a new approach to ML-SSS (Maximum Likelihood Successive State Splitting) algorithm that uses tied-mixture representation of the output probability density function instead of a single Gaussian during the splitting phase of the ML-SSS algorithm. The tied-mixture representation results in a better state split gain, because it is able to(More)
  • 1