Estimation of general identifiable linear dynamic models with an application in speech recognition
2007 (English)In: 2007 IEEE International Conference on Acoustics, Speech, and Signal Processing, Vol IV, Pts 1-3 / [ed] IEEE, 2007, 453-456 p.Conference paper (Refereed)
Although Hidden Markov Models (HMMs) provide a relatively efficient modeling framework for speech recognition, they suffer from several shortcomings which set upper bounds in the performance that can be achieved. Alternatively, linear dynamic models (LDM) can be used to model speech segments. Several implementations of LDM have been proposed in the literature. However, all had a restricted structure to satisfy identifiability constraints. In this paper, we relax all these constraints and use a general, canonical form for a linear state-space system that guarantees identifiability for arbitrary state and observation vector dimensions. For this system,we present a novel, element-wise Maximum Likelihood (ML) estimation method. Classification experiments on the AURORA2 speech database show performance gains compared to HMMs, particularly on highly noisy conditions.
Place, publisher, year, edition, pages
2007. 453-456 p.
, International Conference on Acoustics Speech and Signal Processing (ICASSP), ISSN 1520-6149
Speech Recognition, Modeling, Identification
Signal Processing Computer Engineering
IdentifiersURN: urn:nbn:se:kth:diva-49542DOI: 10.1109/ICASSP.2007.366947ISI: 000248909200114ISBN: 1-4244-0727-3OAI: oai:DiVA.org:kth-49542DiVA: diva2:459789
32nd IEEE International Conference on Acoustics, Speech, and Signal Processing. Honolulu, HI, USA. 15 April 2007 - 20 April 2007
QC 201111302011-11-282011-11-282011-11-30Bibliographically approved