Name: | Description: | Size: | Format: | |
---|---|---|---|---|
In this paper a hierarchical classification of different levels of phonetic information is proposed in order to improve phone recognition. In this paradigm several intermediate classifiers give posterior probability predictions for broad phonetic classes, achieving phone detail in the last layer. Class membership probabilities are weighted and combined in order to get a more robust phoneme prediction. A method for finding the best set of weights is also proposed based on discriminative training in a hybrid MLP/HMM system. Experiments show that the use of broad-class information enhances phone recognition. Relative improvements of 8% in Correctness and 5% in Accuracy were achieved in phoneme recognition on the TIMIT database compared to a baseline system. | 627.59 KB | Adobe PDF |
Advisor(s)
Abstract(s)
In this paper a hierarchical classification of different levels of phonetic information is proposed in order to improve phone recognition. In this paradigm several intermediate classifiers give posterior probability predictions for broad phonetic classes, achieving phone detail in the last layer. Class membership probabilities are weighted and combined in order to get a more robust phoneme prediction. A method for finding the best set of weights is also proposed based on discriminative training in a hybrid MLP/HMM system. Experiments show that the use of broad-class information enhances phone recognition. Relative improvements of 8% in Correctness and 5% in Accuracy were achieved in phoneme recognition on the TIMIT database compared to a baseline system.
Description
17th European Signal Processing Conference, EUSIPCO 2009, 24 August 2009 through 28 August 2009 - Code 91099
Keywords
Hidden Markov models Abstracts Databases Engines
Citation
C. Lopes and F. Perdigão, "A hierarchical broad-class classification to enhance phoneme recognition," 2009 17th European Signal Processing Conference, Glasgow, UK, 2009, pp. 1760-1764.
Publisher
IEEE Canada
CC License
Without CC licence