Posterior-based confidence measures for spoken term detection
Entity
UAM. Departamento de Ingeniería InformáticaPublisher
Institute of Electrical and Electronics EngineersDate
2009Citation
10.1109/ICASSP.2009.4960727
IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2009, IEEE, 2009, 4889 - 4892.
ISSN
1520-6149DOI
10.1109/ICASSP.2009.4960727Funded by
This work used the Edinburgh Compute and Data Facility which is partially supported by eDIKT. Part of this work was funded by the Spanish Ministry of Science and Education (TIN 2005-06885).Editor's Version
http://dx.doi.org/10.1109/ICASSP.2009.4960727Subjects
Confidence measure; MLP; Posterior probabilities; Spoken term detection; Informática; TelecomunicacionesNote
Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. D. Wang, J. Tejedor, J. G. Frankel, S. King, and J. Colás, "Posterior-based confidence measures for spoken term detection", IEEE International Conference on Acoustics, Speech and Signal Processing, 2009. ICASSP 2009, Taipei (China), 2009, pp. 4889 - 4892Rights
© 2009 IEEEAbstract
Confidence measures play a key role in spoken term detection (STD) tasks. The confidence measure expresses the posterior probability of the search term appearing in the detection period, given the speech. Traditional approaches are based on the acoustic and language model scores for candidate detections found using automatic speech recognition, with Bayes' rule being used to compute the desired posterior probability. In this paper, we present a novel direct posterior-based confidence measure which, instead of resorting to the Bayesian formula, calculates posterior probabilities from a multi-layer perceptron (MLP) directly. Compared with traditional Bayesian-based methods, the direct-posterior approach is conceptually and mathematically simpler. Moreover, the MLP-based model does not require assumptions to be made about the acoustic features such as their statistical distribution and the independence of static and dynamic co-efficients. Our experimental results in both English and Spanish demonstrate that the proposed direct posterior-based confidence improves STD performance.
Files in this item
Google Scholar:Wang, Dong
-
Tejedor Noguerales, Javier
-
Frankel, Joe
-
King, Simon
-
Colás Pasamontes, José
This item appears in the following Collection(s)
Related items
Showing items related by title, author, creator and subject.