F. Perdigao

  • Citations Per Year
Learn More
Unlike in data communications, multimedia communications through heterogeneous networks is still far away of being achieved due to lacking of solutions at all protocol layers. The main objective of this research work is to evaluate the performance of a high frequency (HF) wireless network for transporting multimedia services. Beyond of allowing(More)
The challenge of this paper is to extend the concept of discriminative training to a hybrid ANN/HMM phoneme recognition system. The main goal is to improve phoneme accuracy in the aligned output string, instead of in the multi layer perceptron (MLP) output, as usually done. The method uses the difference between the reference and the best acoustic(More)
Future end-to-end network delivering allows us to access anytime and anywhere to the video based multimedia contents. Ionosphere communications is part of the future network convergence and interoperability. Thus, this paper evaluates the significance of MPEG-4 coded speech parameters at a very low bit rate, 2 and 4 kbps transmitted over long distance short(More)
This paper is intended as an overview of the REC project. Hence, its starts with a description of the original goals and the reasons which led us to restructure the proposed workplan. A short summary of the main results in each task is then included, followed by a section on which were, in the opinion of the consortium members, the strongest and weakest(More)
A powerful feature extraction system for noise robust speech recognition was standardized by ETSI. The system was developed for distributed speech recognition (DSR) and includes an advanced front-end (AFE) to be implemented in client terminals, which send the extracted parameters to a remote server that runs a speech recognition engine. In view of the(More)
Future end-to-end network delivering allows us to access to multimedia contents anytime and anywhere. Ionosphere communications is part of the future network convergence and interoperability. This paper deals with digital speech communications at very low bit rates like 2 and 4 kbps transmitted over long distance short wave channels at High Frequencies (HF)(More)
In this paper a new algorithm is proposed for fast discriminative training of hidden Markov models (HMMs) based on minimum classification error (MCE). The algorithm is able to train acoustic models in a few iterations, thus overcoming the slow training speed typical of discriminative training methods based on gradient-descendent. The algorithm tries to(More)
  • 1