abstract = {There is strong neurophysiological evidence sug- gesting that processing of speech signals in the brain happens along parallel paths which encode complementary information in the signal. These parallel streams are organized around a duality of slow vs. fast: Coarse signal dynamics appear to be processed separately from rapidly changingmodulations both in the spectral and temporal dimensions.We adapt such duality in amultistream framework for robust speaker-independent phoneme recognition. The scheme presented here centers around a multi-path bandpass modulation analysis of speech sounds with each streamcovering an entire range of temporal and spectral modulations. By performing bandpass operations along the spectral and temporal dimensions, the proposed scheme avoids the classic feature explosion problem of previous multistream approaches while maintaining the advan- tage of parallelism and localized feature analysis. The proposed architecture results in substantial improvements over standard and state-of-the-art feature schemes for phoneme recognition, particularly in presence of nonstationary noise, reverberation and channel distortions.},
author = {Nemala, Sridhar Krishna and Patil, Kailash and Elhilali, Mounya},
doi = {10.1109/TASL.2012.2219526},
isbn = {1558-7916},
issn = {1558-7916},
journal = {IEEE Transactions on Audio, Speech, and Language Processing},
keywords = {Auditory cortex,automatic speech recognition (ASR),modulation,multistream,speech parameterization},
number = {2},
pages = {416--426},
title = {{A Multistream Feature Framework Based on Bandpass Modulation Filtering for Robust Speech Recognition}},
url = {http://ieeexplore.ieee.org/document/6305465/},
volume = {21},
year = {2013}