Independent component analysisapplied to feature extraction forrobust automatic speechrecognition
Ilyas Potamitis, Nikos Fakotakis, G. Kokkinakis · Electronics Letters · 2000
The authors explore independent component analysis (ICA) as a statistical technique for deriving suitable data-driven representational bases for the projection of spectra and cepstra in the context of automatic speech recognition (ASR). Based on the close link between the independent mechanisms of speech variability and the concept of statistical independence they derive a new feature transformation that effects consistent improvement in recognition performance.