Developmental audio pattern recognition
Riken Brain Science Institute, 2009
Since Jan. 2009, I've been lucky to collaborate with biologists in Kazuo Okanoya's laboratory (Riken Brain Science Institute, Saitama, Japan), who study the evolution and development of animal acoustic communication. This means a wealth (read: Terabytes) of formidable birds, rats, monkey etc. audio recordings which I'm subjecting to computer analysis. Traditional techniques don't necessarily apply well, as they rely on the availability of a stable training dataset; here, data statistics changes everyday, as young animals develop more and more sophisticated and adult acoustic abilities.
Joint work with: Kazuo Okanoya (Laboratory head, project leader, JST/ERATO and Riken Brain Science Institute), Kazutoshi Sasahara (UCLA), Yulri Nonaka (Riken Brain Science Institute), Kentaro Katahira (JST/ERATO)
Some papers:
Nonaka, Yulri and Aucouturier, JJ and Katahira, Kentaro and Okanoya, Kazuo Developmental Differentiation in Infant Cry: Private and Universal Features Reflect Protolinguistic properties. EvoLANG 2010 (submitted).
Aucouturier, J.-J. and Okanoya, K. Different time scales for statistical and sequence learning in bird song development: A quantification by acoustic analysis. Journal of the Acoustical Society of America (submitted, 2009).
Aucouturier, J.-J. and Nonaka, Yulri and Katahira, Kentaro and Okanoya, K.
Automatic identification of expiratory and inspiratory sounds in baby cry audio recordings, Journal of Voice (submitted, 2009)
