Improved hidden Markov modeling of phonemes for continuous speech recognition
- 24 March 2005
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 9, 21-24
- https://doi.org/10.1109/icassp.1984.1172751
Abstract
This paper discusses the use of the Hidden Markov Model (HMM) in phonetic recognition. In particular, we present improvements that deal with the problems of modeling the effect of phonetic context and the problem of robust pdf estimation. The effect of phonetic context is taken into account by conditioning the probability density functions (pdfs) of the acoustic parameters on the adjacent phonemes, only to the extent that there are sufficient tokens of the phoneme in that context. This partial conditioning is achieved by combining the conditioned and unconditioned pdfs models with weights that depend on the confidence in each pdf estimate. This combination is shown to result in better performance than either model by itself. We also show that it is possible to obtain the computational advantages of using discrete probability densities without the usual requirement for large amounts of training data.Keywords
This publication has 12 references indexed in Scilit:
- A preliminary design of a phonetic vocoder based on a diphone modelPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- Experiments in syllable-based recognition of continuous speechPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- Advanced acoustic techniques in automatic speech understandingPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- Automatic recognition of continuously spoken sentences from a finite state grammerPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- A preliminary study on the use of demisyllables in automatic speech recognitionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- A Maximum Likelihood Approach to Continuous Speech RecognitionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1983
- The 1976 modular acoustic processor(MAP)IEEE Transactions on Acoustics, Speech, and Signal Processing, 1977
- Design of a linguistic statistical decoder for the recognition of continuous speechIEEE Transactions on Information Theory, 1975
- The DRAGON system--An overviewIEEE Transactions on Acoustics, Speech, and Signal Processing, 1975
- An inequality with applications to statistical estimation for probabilistic functions of Markov processes and to a model for ecologyBulletin of the American Mathematical Society, 1967