PMID- 26388721
OWN - NLM
STAT- PubMed-not-MEDLINE
DCOM- 20150921
LR  - 20231104
IS  - 1662-4548 (Print)
IS  - 1662-453X (Electronic)
IS  - 1662-453X (Linking)
VI  - 9
DP  - 2015
TI  - Sound stream segregation: a neuromorphic approach to solve the "cocktail party 
      problem" in real-time.
PG  - 309
LID - 10.3389/fnins.2015.00309 [doi]
LID - 309
AB  - The human auditory system has the ability to segregate complex auditory scenes 
      into a foreground component and a background, allowing us to listen to specific 
      speech sounds from a mixture of sounds. Selective attention plays a crucial role 
      in this process, colloquially known as the "cocktail party effect." It has not 
      been possible to build a machine that can emulate this human ability in 
      real-time. Here, we have developed a framework for the implementation of a 
      neuromorphic sound segregation algorithm in a Field Programmable Gate Array 
      (FPGA). This algorithm is based on the principles of temporal coherence and uses 
      an attention signal to separate a target sound stream from background noise. 
      Temporal coherence implies that auditory features belonging to the same sound 
      source are coherently modulated and evoke highly correlated neural response 
      patterns. The basis for this form of sound segregation is that responses from 
      pairs of channels that are strongly positively correlated belong to the same 
      stream, while channels that are uncorrelated or anti-correlated belong to 
      different streams. In our framework, we have used a neuromorphic cochlea as a 
      frontend sound analyser to extract spatial information of the sound input, which 
      then passes through band pass filters that extract the sound envelope at various 
      modulation rates. Further stages include feature extraction and mask generation, 
      which is finally used to reconstruct the targeted sound. Using sample tonal and 
      speech mixtures, we show that our FPGA architecture is able to segregate sound 
      sources in real-time. The accuracy of segregation is indicated by the high 
      signal-to-noise ratio (SNR) of the segregated stream (90, 77, and 55 dB for 
      simple tone, complex tone, and speech, respectively) as compared to the SNR of 
      the mixture waveform (0 dB). This system may be easily extended for the 
      segregation of complex speech signals, and may thus find various applications in 
      electronic devices such as for sound segregation and speech recognition.
FAU - Thakur, Chetan Singh
AU  - Thakur CS
AD  - Biomedical Engineering and Neuroscience, The MARCS Institute, University of 
      Western Sydney Sydney, NSW, Australia.
FAU - Wang, Runchun M
AU  - Wang RM
AD  - Biomedical Engineering and Neuroscience, The MARCS Institute, University of 
      Western Sydney Sydney, NSW, Australia.
FAU - Afshar, Saeed
AU  - Afshar S
AD  - Biomedical Engineering and Neuroscience, The MARCS Institute, University of 
      Western Sydney Sydney, NSW, Australia.
FAU - Hamilton, Tara J
AU  - Hamilton TJ
AD  - Biomedical Engineering and Neuroscience, The MARCS Institute, University of 
      Western Sydney Sydney, NSW, Australia.
FAU - Tapson, Jonathan C
AU  - Tapson JC
AD  - Biomedical Engineering and Neuroscience, The MARCS Institute, University of 
      Western Sydney Sydney, NSW, Australia.
FAU - Shamma, Shihab A
AU  - Shamma SA
AD  - Department of Electrical and Computer Engineering and Institute for Systems 
      Research, University of Maryland College Park, MD, USA.
FAU - van Schaik, Andre
AU  - van Schaik A
AD  - Biomedical Engineering and Neuroscience, The MARCS Institute, University of 
      Western Sydney Sydney, NSW, Australia.
LA  - eng
GR  - R01 DC007657/DC/NIDCD NIH HHS/United States
PT  - Journal Article
DEP - 20150902
PL  - Switzerland
TA  - Front Neurosci
JT  - Frontiers in neuroscience
JID - 101478481
PMC - PMC4557082
OTO - NOTNLM
OT  - FPGA
OT  - cochlea
OT  - cocktail party problem
OT  - machine-based speech recognition
OT  - temporal coherence
EDAT- 2015/09/22 06:00
MHDA- 2015/09/22 06:01
PMCR- 2015/01/01
CRDT- 2015/09/22 06:00
PHST- 2015/05/19 00:00 [received]
PHST- 2015/08/18 00:00 [accepted]
PHST- 2015/09/22 06:00 [entrez]
PHST- 2015/09/22 06:00 [pubmed]
PHST- 2015/09/22 06:01 [medline]
PHST- 2015/01/01 00:00 [pmc-release]
AID - 10.3389/fnins.2015.00309 [doi]
PST - epublish
SO  - Front Neurosci. 2015 Sep 2;9:309. doi: 10.3389/fnins.2015.00309. eCollection 
      2015.