Title :
Analysis of phone posterior feature space exploiting class-specific sparsity and MLP-based similarity measure
Author :
Asaei, Afsaneh ; Picart, Benjamin ; Bourlard, Hervé
Author_Institution :
IDIAP Res. Inst., Martigny, Switzerland
Abstract :
Class posterior distributions have recently been used quite successfully in Automatic Speech Recognition (ASR), either for frame or phone level classification or as acoustic features, which can be further exploited (usually after some “ad hoc” transformations) in different classifiers (e.g., in Gaussian Mixture based HMMs). In the present paper, we show preliminary results showing that it may be possible to perform speech recognition without explicit subword unit (phone) classification or likelihood estimation, simply answering the question whether two acoustic (posterior) vectors belong to the same subword unit class or not. In this paper, we first exhibit specific properties of the posterior acoustic space before showing how those properties can be exploited to reach very high performance in deciding (based on an appropriate, trained, distance metric, and hypothesis testing approaches) whether two posterior vectors belong to the same class or not. Performance as high as 90% correct decision rates are reported on the TIMIT database, before reporting kNN phone classification rates.
Keywords :
multilayer perceptrons; pattern classification; speech recognition; MLP-based similarity measure; TIMIT database; acoustic vectors; automatic speech recognition; class posterior distributions; class-specific sparsity; kNN phone classification rates; phone posterior feature space; Acoustic testing; Automatic speech recognition; Extraterrestrial measurements; Feature extraction; Hidden Markov models; Mel frequency cepstral coefficient; Multilayer perceptrons; Spatial databases; Speech analysis; Speech recognition; Posterior feature space; kNN classifier; posterior space properties; posterior-based metrics;
Conference_Titel :
Acoustics Speech and Signal Processing (ICASSP), 2010 IEEE International Conference on
Conference_Location :
Dallas, TX
Print_ISBN :
978-1-4244-4295-9
Electronic_ISBN :
1520-6149
DOI :
10.1109/ICASSP.2010.5495121