A Cortically-Inspired Model for Bioacoustics Recognition [chapter]

Linda Main, John Thornton
2015 Lecture Notes in Computer Science  
Wavelet transforms have shown superior performance in bioacoustic recognition tasks compared to the more commonly used Mel-Frequency Cepstral Coefficients, and offer the ability to more closely model the frequency response behaviour of the basilar membrane within the cochlea. In this paper we evaluate a gammatone wavelet as a preprocessor for the Hierarchical Temporal Memory (HTM) model of the neocortex, as part of the broader development of a biologically motivated approach to sound
more » ... . Specifically, we implement a gammatone/equivalent rectangular bandwidth wavelet transform and apply it, in conjunction with the HTM's spatial pooler, to recognise frog calls, bird songs and insect sounds. We demonstrate the improved performance of wavelets for feature detection and the potential viability of using HTM for bioacoustic recognition. Our classification accuracy of 99.5% in detecting insect sounds and 96.3% in detecting frog calls are significant improvements on results previously published for the same datasets.
doi:10.1007/978-3-319-26561-2_42 fatcat:7ziv5kf3pvgi7kungkv2bly4ni