Unsupervised spoken keyword spotting via segmental DTW on Gaussian posteriorgrams
Authors:
Top Cited Papers
- 1 December 2009
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 398-403
Abstract
In this paper, we present an unsupervised learning framework to address the problem of detecting spoken keywords. Without any transcription information, a Gaussian Mixture Model is trained to label speech frames with a Gaussian posteriorgram. Given one or more spoken examples of a keyword, we use segmental dynamic time warping to compare the Gaussian posteriorgrams between keyword samples and test utterances. The keyword detection result is then obtained by ranking the distortion scores of all the test utterances. We examine the TIMIT corpus as a development set to tune the parameters in our system, and the MIT Lecture corpus for more substantial evaluation. The results demonstrate the viability and effectiveness of our unsupervised learning framework on the keyword spotting task.Keywords
This publication has 18 references indexed in Scilit:
- Unsupervised Pattern Discovery in SpeechIEEE Transactions on Audio, Speech, and Language Processing, 2007
- A Novel Instance Matching Based Unsupervised Keyword Spotting System2007
- Improvements in phone based audio search via constrained match with high order confusion estimates2007
- Keyword Spotting of Arbitrary Words Using Minimal Speech Resources2006
- A probabilistic framework for segment-based speech recognitionComputer Speech & Language, 2003
- Maximum likelihood successive state splitting2002
- A hidden Markov model based keyword recognition system2002
- LVCSR log-likelihood ratio scoring for keyword spotting2002
- Lightly supervised and unsupervised acoustic model trainingComputer Speech & Language, 2002
- Dynamic programming algorithm optimization for spoken word recognitionIEEE Transactions on Acoustics, Speech, and Signal Processing, 1978