By Topic

Keyword Spotting using Vowel Onset Point, Vector Quantization and Hidden Markov Modeling Based techniques

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

3 Author(s)
Reddy, B.V.S. ; Dept. of Electron. & Commun. Eng., Indian Inst. of Technol., Guwahati ; Rao, K.V. ; Prasanna, S.

This work demonstrates the development of Keyword Spotting (KWS) system using Vowel Onset Point (VOP), Vector Quantization (VQ) and Hidden Markov Model(HMM) based techniques. The goal of KWS system is to spot the keywords present in the test speech signal, while neglecting rest of the words. In this work, first independent KWS systems will be developed using VOP, VQ and HMM techniques. Each of these methods involve different techniques and hence it may be possible to combine them for achieving higher performance. In the next step, KWS system is also developed by combining HMM and VQ (HMM-VQ) and also HMM and VOP (HMM-VOP) based KWS systems. The performance measured in terms of Figure Of Merit (FOM) for HMM, VQ and VOP are 53.32, 22.41 and 26.95, respectively. The FOM of combinations HMM-VQ and HMM-VOP are 57.18 and 60.62, respectively. The significantly improved performance in the combined systems demonstrate the complementary nature of keyword information exploited by each of the independent systems.

Published in:

TENCON 2008 - 2008 IEEE Region 10 Conference

Date of Conference:

19-21 Nov. 2008