Emulating Temporal Receptive Fields of Auditory Mid-Brain Neurons for Automatic Speech Recognition

Sivaram, G. S. V. S.; Hermansky, Hynek

Emulating Temporal Receptive Fields of Auditory Mid-Brain Neurons for Automatic Speech Recognition

This paper proposes modifications to the Multi-resolution RASTA (MRASTA) feature extraction technique for the automatic speech recognition (ASR). By emulating asymmetries of the temporal receptive field (TRF) profiles of auditory mid-brain neurons, we obtain more than $13\% $ relative improvement in word error rate on OGI-Digits database. Experiments on TIMIT database confirm that proposed modifications are indeed useful.

Presented at:

Proc. 16th European Signal Processing Conference (EUSIPCO)

Year:

2008

Note:

IDIAP-RR 08-24

Laboratories:

LIDIAP

Record appears in:

Scientific production and competences > STI - School of Engineering > IEL - Institute of Electrical Engineering > LIDIAP - L'IDIAP Laboratory
Scientific production and competences > Euler Center for Signal Processing
Publicly available
Conference Papers
Work produced at EPFL
Published