Emulating Temporal Receptive Fields of Higher Level Auditory Neurons for ASR

• Presented at: 11th International Conference on Text, Speech and Dialogue, Brno, CZECH REPUBLIC, Sep 08-12, 2008
• Published in: Text, Speech And Dialogue, Proceedings, vol. 5246, p. 509-516
• Series: Lecture Notes In Artificial Intelligence
• Springer-Verlag New York, Ms Ingrid Cunningham, 175 Fifth Ave, New York, Ny 10010 Usa, 2008

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 higher level auditory neurons, we obtain more than 11.4% relative improvement in word error rate on OGI-Digits database. Experiments on TIMIT database confirm that proposed modifications are indeed useful.

Keywords: Feature extraction ; auditory neurons and speech recognition ; Cortex

Reference

• EPFL-CONF-160530
• View record in Web of Science

Record created on 2010-11-30, modified on 2017-05-10