Comparison of Different Approaches for Speech Recognition in Hands-free Mode

Conference: Sprachkommunikation - Beiträge zur 10. ITG-Fachtagung
09/26/2012 - 09/28/2012 at Braunschweig, Deutschland

Proceedings: Sprachkommunikation

Pages: 4Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Hirsch, Hans-Günter (Institute for Pattern Recognition, Niederrhein University of Applied Sciences, Krefeld, Germany)
Ganapathy, Sriram; Hermansky, Hynek (Center for Language and Speech Processing, Johns Hopkins University, Baltimore, USA)

To improve speech recognition in case of a hands-free speech input we apply the signal processing technique known as frequency domain linear prediction (FDLP). By analyzing the effects of reverberation we prove that this method is well suited to create robust acoustic features for this mode of speech input. Furthermore, we compare the efficiency of this robust feature extraction scheme with the alternative approach of adapting the Hidden Markov Models (HMMs). We especially investigate the influence of a varying distance between the speaker and the microphone in a reverberant room.