Sixth International Conference on Spoken Language Processing
(ICSLP 2000)

Beijing, China
October 16-20, 2000

Spectral and Cepstral Projection Bases Constructed by Independent Component Analysis

Ilyas Potamitis, Nikos Fanotakis, George Kokkinakis

Wire Communications Laboratory, Univ. of Patras, Greece

The present paper addresses the question of the efficiency of Independent Coitiponent Analysis (ICA) as a statistical process for deriving optimal representational bases for the projection of spectrum and cepstrum in the context of Automatic Speech Recognition (ASR). Several decorrelation strategies have been applied on the log-spectrum and cepstrum to fulfill the practical need of a diagonal covariance HMM for uncorrelated features. In our work we question the optimality of a fixed decorrelation strategy as DCT and follow an emerging trend in ASR that designs projection bases based on the statistics of speech. We differentiate our approach from the second order statistics of Discrete Cosine Transform (DCT), Linear Discriminatio Analysis (LDA) and Principal Component Analysis (PCA) by proposing an alternative data-driven approach based on higher Order Statistics.

Full Paper

Bibliographic reference.  Potamitis, Ilyas / Fanotakis, Nikos / Kokkinakis, George (2000): "Spectral and cepstral projection bases constructed by independent component analysis", In ICSLP-2000, vol.3, 63-66.