Spectral kurtosis: from definition to application

The objective of this paper is to present the Spectral Kurtosis (SK) as a spectral analysis tool complementary to the classical power spectrum density (PSD). Theoretical definition of the SK is firstly justified, followed by a study of its behavior with respect to the spectral components of signals. An unbiased estimator is thus proposed, with its exact variance, and its signal to noise ratio sensivity. Finally, results are illustrated by applying SK and PSD on synthetic and real signals.

[1]  Ch. Serviere,et al.  Feasibility of source separation in frequency domain , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[2]  Jean-Louis Lacoume,et al.  Statistics for complex variables and signals - Part I: Variables , 1996, Signal Process..

[3]  Jean-Louis Lacoume,et al.  Blind separation of wide-band sources: Application to rotating machine signals , 1996, 1996 8th European Signal Processing Conference (EUSIPCO 1996).

[4]  Giorgio Tacconi,et al.  Filtering of randomly occurring signals by kurtosis in the frequency domain , 1994, Proceedings of the 12th IAPR International Conference on Pattern Recognition, Vol. 2 - Conference B: Computer Vision & Image Processing. (Cat. No.94CH3440-5).