All

What are you looking for?

All
Projects
Results
Organizations

Quick search

  • Projects supported by TA ČR
  • Excellent projects
  • Projects with the highest public support
  • Current projects

Smart search

  • That is how I find a specific +word
  • That is how I leave the -word out of the results
  • “That is how I can find the whole phrase”

Fast Approximate Spoken Term Detection from Sequence of Phonemes

The result's identifiers

  • Result code in IS VaVaI

    <a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F08%3APU80187" target="_blank" >RIV/00216305:26230/08:PU80187 - isvavai.cz</a>

  • Result on the web

  • DOI - Digital Object Identifier

Alternative languages

  • Result language

    angličtina

  • Original language name

    Fast Approximate Spoken Term Detection from Sequence of Phonemes

  • Original language description

    We investigate the detection of spoken terms in conversa-<br> tional speech using phoneme recognition with the objective<br> of achieving smaller index size as well as faster search speed.<br> Speech is processed and indexed as a sequence of one best<br>phoneme sequence. We propose the use of a probabilistic<br> pronunciation model for the search term to compensate for<br> the errors in the recognition of phonemes. This model is de-<br> rived using the pronunciation of the word and the phoneme<br> confusion matrix. Experiments are performed on the con-<br> versational telephone speech database distributed by NIST<br> for the 2006 spoken term detection. We achieve about 1500<br> times smaller index size and 14 times faster search speed<br> compared tothe system using phoneme lattices, at the cost<br> of relatively lower detection performance.

  • Czech name

    Fast Approximate Spoken Term Detection from Sequence of Phonemes

  • Czech description

    We investigate the detection of spoken terms in conversa-<br>tional speech using phoneme recognition with the objective<br>of achieving smaller index size as well as faster search speed.<br>Speech is processed and indexed as a sequence of one best<br>phoneme sequence. We propose the use of a probabilistic<br>pronunciation model for the search term to compensate for<br>the errors in the recognition of phonemes. This model is de-<br>rived using the pronunciation of the word and the phoneme<br>confusion matrix. Experiments are performed on the con-<br>versational telephone speech database distributed by NIST<br>for the 2006 spoken term detection. We achieve about 1500<br>times smaller index size and 14 times faster search speed<br>compared to the system using phoneme lattices, at the cost<br>of relatively lower detection performance.<br><br>

Classification

  • Type

    D - Article in proceedings

  • CEP classification

    JC - Computer hardware and software

  • OECD FORD branch

Result continuities

  • Project

  • Continuities

    Z - Vyzkumny zamer (s odkazem do CEZ)

Others

  • Publication year

    2008

  • Confidentiality

    S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Data specific for result type

  • Article name in the collection

    The 31st Annual International ACM SIGIR Conference 20-24 July 2008, Singapore

  • ISBN

    978-90-365-2697-5

  • ISSN

  • e-ISSN

  • Number of pages

    8

  • Pages from-to

  • Publisher name

    Association for Computing Machinery

  • Place of publication

    Singapore

  • Event location

    Singapur

  • Event date

    Jul 20, 2008

  • Type of event by nationality

    WRD - Celosvětová akce

  • UT code for WoS article