ICA based on the data asymmetry

TL;DR

This paper introduces a new ICA method based on the Split Gaussian distribution, effectively handling asymmetric data and outperforming classical methods in non-symmetric scenarios like image color distributions.

Contribution

The paper presents a novel ICA approach utilizing the Split Gaussian distribution, addressing the gap in handling asymmetric data more effectively.

Findings

Outperforms classical ICA methods on asymmetric data

Particularly effective for image color distribution analysis

Demonstrates improved independence separation in non-symmetric cases

Abstract

Independent Component Analysis (ICA) - one of the basic tools in data analysis - aims to find a coordinate system in which the components of the data are independent. Most of existing methods are based on the minimization of the function of fourth-order moment (kurtosis). Skewness (third-order moment) has received much less attention. In this paper we present a competitive approach to ICA based on the Split Gaussian distribution, which is well adapted to asymmetric data. Consequently, we obtain a method which works better than the classical approaches, especially in the case when the underlying density is not symmetric, which is a typical situation in the color distribution in images.