Parameter estimation of Oscillating Gaussian functions using the scaled reassigned spectrogram
Research output: Contribution to journal › Article
In this paper we suggest an algorithm for estimation of the parameters detailing Oscillating Gaussian functions. The different components of the signal are first detected in the spectrogram. After this we exploit the fact that a Gaussian function may be perfectly reassigned into one single point given a correct scaling factor, where this scaling factor is a function of the unknown shape parameter of the Gaussian function. The scaled reassignment of the spectrogram is performed using a set of candidate scaling factors and the local Renyi entropy is used to measure the concentration of each component using every candidate scaling factor. The estimates are refined by using non-linear least squares. The algorithm is evaluated on both simulated and real data.
|Research areas and keywords||
Subject classification (UKÄ) – MANDATORY
|Number of pages||13|
|State||Published - 2018 Sep 1|