A Numerical Integral Algorithm Based on the CAPSO to Improve the Estimation for the Parameters of the Homodyned-K Distribution

Abstract

The homodyned-K (HK) distribution is a widely used statistical model, whose parameters have different physical-meanings for tissue characterization. In the present study, the maximum likelihood estimation (MLE) method based on the Newton-Raphson algorithm is proposed to estimate the HK parameters solely. For improving the accuracy and convergence of the MLE, the cloud adaptive particle swarm optimization (CAPSO) algorithm is proposed for the integral calculation of the probability density function (PDF) of the HK distribution. In the experiments, sets of samples satisfying the HK distribution are generated, and then the parameters are estimated by the proposed CAPSO-based MLE method. The statistics of estimation errors are calculated, and compared with the results based on the mean intensity and X- and U-statistics (XU) method, which is the latest one based on moment estimation. Experimental results show that the proposed method can solely estimate the HK parameters with a small error level, which means a further practical value in ultrasonic applications.