Ultimate Trends in Integrated Systems to Enhance Automatic Speech Recognition Performance

  title={Ultimate Trends in Integrated Systems to Enhance Automatic Speech Recognition Performance},
  author={Carolina Palma Dur{\'a}n},
  • C. Durán
  • Published 1 November 2008
  • Computer Science
An automatic speech recognition (ASR) system can be defined as a mechanism capable of decoding the signal produced in the vocal and nasal tracts of a human speaker into the sequence of linguistic units contained in the message that the speaker wants to communicate (Peinado & Segura, 2006). The final goal of ASR is the man–machine communication. This natural way of interaction has found many applications because of the fast development of different hardware and software technologies. The most… 
and Intelligent Sensors 2012 ( IRIS 2012 ) Isolated Word Command Recognition for Robot Navigation
This project is demonstrating Robot Navigation Control using Isolated Word Recognition by using MATLAB programming code to recognize the word Normalization, Power density spectrum and voice fingerprint method are used and implemented by MATLAB software code.
Interactive Dictionary for Visually Impaired
This paper is presenting interactive learning tools for visually impaired people, an audio dictionary which gives synonyms of words spoken by the user through microphone, step ahead of the speech synthesis technology where the speech is not only synthesized but also gives thesynonyms of the words in the form of audio output.


Speech translation enhanced automatic speech recognition
This work presents an iterative approach for ASR improvement and outperform the baseline system by a relative word error rate reduction and shows how multiple target languages, as for example provided by different simultaneous translators during European Parliament debates, can be incorporated into the system design.
Integrated System Approach for the Automatic Speech Recognition using Linear predict Coding and Neural Networks
An automatic speech recognition approach to identify initially four voice words using the energy of the signal through of LPC (linear predict coding) and finally neural networks as recognition and classification techniques of speech parameters.
Isolated word recognition of deaf speech using artificial neural networks
Describes the development and implementation of a prototype speech recognition system for carrying out isolated word recognition of deaf speech using a combination of artificial neural networks and conventional signal processing techniques.
On design and implementation of an embedded automatic speech recognition system
Both the hardware and the software have been designed concurrently, with a view to achieve high-speed recognition with maximum accuracy in minimum power and making the device portable.
Learning name pronunciations in automatic speech recognition systems
This paper describes a data-driven method that learns proper name pronunciations from audio samples of these words and shows that the proposed algorithm reduces the call routing error rate by 40% compared to a reference letter-to-phone pronunciation engine.
Statistical voice activity detection using a multiple observation likelihood ratio test
This letter presents a new voice activity detector (VAD) for improving speech detection robustness in noisy environments and the performance of speech recognition systems using an optimum likelihood ratio test (LRT) involving multiple and independent observations.
Robust Technologies towards Automatic Speech Recognition in Car Noise Environments
This paper proposes to use an immunity learning scheme, in which pre-recorded car noises are artificially added to clean training utterances with different signal-to-noise ratios (SNR) to imitate the in-car environments to improve robust automatic speech recognition in car noise environments.
The Use of Multilayer Perceptrons in Isolated Word Recognition
The technique called Trace Segmentation is explored in order to fit the lengths of both input layer and utterances, achieving results comparable to or higher than those obtained with Hidden Markov Models or more conventional and expensive Multilayer Perceptrons when applied to the same task.