FFT based automatic species identification improvement with 4-layer neural network

In this paper, an automatic species identification system has been developed. Recoded data was segmented, processed, features taken out, and identified by an automatic operation. A feature quantity method based on FFT with derivative of frequency band power making use of 4-layer neural network is proposed. Comparison of the results with the 4-layer neural network has been performed on wild bird species identification based on sound data which has proved promising.

[1]  Rong Sun,et al.  Nocturnal wild bird species identification by sound information using wavelet , 2012, 2012 International Conference on Wavelet Active Media Technology and Information Processing (ICWAMTIP).

[2]  Juha T. Tanttu,et al.  Wavelets in Recognition of Bird Sounds , 2007, EURASIP J. Adv. Signal Process..

[3]  Hani G. Melhem,et al.  Fourier and wavelet analyses for fatigue assessment of concrete beams , 2003 .

[4]  H. C. Card,et al.  Birdsong recognition using backpropagation and multivariate statistics , 1997, IEEE Trans. Signal Process..

[5]  Misha Pavel,et al.  On the relative importance of various components of the modulation spectrum for automatic speech recognition , 1999, Speech Commun..

[6]  Tony R. Martinez,et al.  Digital Neural Networks , 1988, Proceedings of the 1988 IEEE International Conference on Systems, Man, and Cybernetics.

[7]  Takayuki Arai,et al.  Robust Automatic Speech Recognition Emphasizing Important Modulation Spectrum , 2001 .

[8]  S.-A. Selouani,et al.  Automatic birdsong recognition based on autoregressive time-delay neural networks , 2005, 2005 ICSC Congress on Computational Intelligence Methods and Applications.

[9]  H. C. Card,et al.  A comparison of backpropagation and statistical classifiers for bird identification , 1997, Proceedings of International Conference on Neural Networks (ICNN'97).