A constrained sequential EM algorithm for speech enhancement
SCIE
SCOPUS
- Title
- A constrained sequential EM algorithm for speech enhancement
- Authors
- Park, S; Choi, S
- Date Issued
- 2008-11
- Publisher
- PERGAMON-ELSEVIER SCIENCE LTD
- Abstract
- Speech enhancement is a fundamental problem, the goal of which is to estimate clean speech s(t), given a noise-contaminated signal s(t) + n(t), where n(t) is white or colored noise, This task call be viewed as a probabilistic inference problem which involves estimating the posterior distribution of hidden clean speech, given a noisy observation. Kalman filter is a representative method but is restricted to Gaussian distributions only. We consider the generalized auto-regressive (GAR) model in order to Capture the non-Gaussian characteristics of speech. Then we present a constrained sequential EM algorithm where Rao-Blackwellized particle filters (RBPFs) are used in the E-step and model parameters are updated in a sequential manner in the M-step under positivity constraints for noise variance parameters. Numerical experiments confirm the high performance of our proposed method, compared to Kalman filter-based methods, in the task of sequential speech enhancement. (C) 2008 Elsevier Ltd. All rights reserved.
- Keywords
- Expectation maximization (EM); Generalized auto-regressive model; Generalized exponential density; Kalman filter; Rao-Blackwellized particle filter; Speech enhancement; SIGNALS; NOISE
- URI
- https://oasis.postech.ac.kr/handle/2014.oak/22360
- DOI
- 10.1016/j.neunet.2008.03.001
- ISSN
- 0893-6080
- Article Type
- Article
- Citation
- NEURAL NETWORKS, vol. 21, no. 9, page. 1401 - 1409, 2008-11
- Files in This Item:
- There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.