A Rich Distribution Family for Blind Biological Signal Separation

摘要

The biological signals measured by multi-sensors are always the mixtures of several independent sources. Therefore, it is necessary to separate them from each other for clinical diagnosis. One of the most challenging problem in signal separation is how to estimate the most appropriate distribution for biological signal data. In this paper, we propose a finite mixture model of modified generalized gamma distributions (GΓD) for robust blind signal separation and data modeling. In a first part of the work we discuss the independent component analysis model, we propose a derivation of the Maximum-Likelihood estimation of the parameters of the new mixture model and we propose an information-theory based approach for the selection of the number of classes. In a second part, we propose some applications relating to blind biosignal separation to measure the performance of the new model in bio signal data modeling.