Conference paper

SZŐKE Igor, GRÉZL František, ČERNOCKÝ Jan and FAPŠO Michal. Acoustic keyword spotter - optimization from end-user perspective. In: Proceedings of the 2010 IEEE Spoken Language Technology Workshop. Berkeley, California: IEEE Signal Processing Society, 2010, pp. 177-181. ISBN 978-1-4244-7902-3.
Publication language:english
Original title:Acoustic keyword spotter - optimization from end-user perspective
Title (cs):Akustický vyhledávač klíčových slov -optimalizace z perspektivy koncového uživatele
Proceedings:Proceedings of the 2010 IEEE Spoken Language Technology Workshop
Conference:IEEE Workshop on Spoken Language Technology
Series:IEEE Catalog Number: CFP 10SLT-USB
Place:Berkeley, California, US
Publisher:IEEE Signal Processing Society
keyword spotting, spoken term detection, neural networks, calibration
This paper is on acoustic keyword spotting. It presents several steps that have to be done to obtain a usable acoustic keyword spotting system. The novelty of the system is in the calibration.
The paper deals with the development of acoustic keyword spotter (KWS) meeting requirements of a real user from the security community. While the basic scheme of the KWS is relatively standard, it uses novel features derived by a hierarchy of neural networks, and score normalization trained to maximize a user-like evaluation metric. The results are reported on a selection of Czech conversational telephone speech (CTS), radio and read data.
   author = {Igor Sz{\H{o}}ke and Franti{\v{s}}ek Gr{\'{e}}zl and Jan
	{\v{C}}ernock{\'{y}} and Michal Fap{\v{s}}o},
   title = {Acoustic keyword spotter - optimization from end-user
   pages = {177--181},
   booktitle = {Proceedings of the 2010 IEEE Spoken Language Technology
   series = {IEEE Catalog Number: CFP 10SLT-USB},
   year = {2010},
   location = {Berkeley, California, US},
   publisher = {IEEE Signal Processing Society},
   ISBN = {978-1-4244-7902-3},
   language = {english},
   url = {}

Your IPv4 address:
Switch to IPv6 connection

DNSSEC [dnssec]