Feature extraction in the Hankel transform domain

Parameter estimation for the class of compound Gaussian random variables is considered in the more natural domain of the Hankel transform, where the expression of the probability density function appears generally more manageable. The estimation algorithm, based on the minimization of the integrated mean squared error between empirical and theoretical Hankel transform, has been tested for the case of K-distributed data using Monte Carlo simulation and some guidelines for the algorithm setup are derived.