Paper Info
Title
Recurrent Timing Neural Networks for Joint F0-Localisation Based Speech Separation
Abstract
A novel extension to recurrent timing neural networks (RTNNs) is proposed which allows such networks to exploit a joint interaural time difference-fundamental frequency (ITD-F0) auditory cue as opposed to F0 only. This extension involves coupling a second layer of coincidence detectors to a two-dimensional RTNN. The coincidence detectors are tuned to particular ITDs and each feeds excitation to a column in the RTNN. Thus, one axis of the RTNN represents F0 and the other ITD. The resulting behaviour allows sources to be segregated on the basis of their separation in ITD-F0 space. Furthermore, all grouping and segregation activity proceeds within individual frequency channels without recourse to across channel estimates of F0 or ITD that are commonly used in auditory scene analysis approaches. The system has been evaluated using a source separation task operating on spatialised speech signals.
Year
DOI
Venue
2007
10.1109/ICASSP.2007.366640
Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference
Keywords
Field
DocType
blind source separation,recurrent neural nets,speech processing,auditory scene analysis approach,channel estimation,coincidence detectors,frequency channels,joint F0-localisation,joint interaural time difference-fundamental frequency,recurrent timing neural networks,segregation activity,speech separation,Auditory system,Neural network architecture,Speech enhancement,Speech processing
Speech enhancement,Speech processing,Auditory scene analysis,Pattern recognition,Computer science,Recurrent neural network,Auditory system,Speech recognition,Artificial intelligence,Artificial neural network,Blind signal separation,Source separation
Conference
Volume
ISSN
ISBN
1
1520-6149 E-ISBN : 1-4244-0728-1
1-4244-0728-1
Citations 
PageRank 
References 
4
0.58
4
Authors
2
Name
Order
Citations
PageRank
Stuart N. Wrigley118120.56
Guy J. Brown276097.54