Blind Source Separation in Perspective of ICA Algorithms: A Review

For many scenarios, the world around us is a never-ending cacophony of signals from various sources, where distinct stimuli do not really mix properly. Separating blind sources is a method of extracting sources from signal mixes without knowing the sources or mechanism of the mixing beforehand. Principal Component Analysis along with Independent Component Analysis and Non-negative Matrix Factorization are frequently used methods for Blind source separation. ICA, a blind statistical methodology, has grabbed the attention of scientists and engineers for numerous applications in signal processing for extracting additive subcomponents from a multivariate signal. This paper attempts to explore blind source separation methods and compare and analyze different algorithms based on a case study. This study also posits a review of different ICA algorithms. A systematic approach for literature review is carried out on the reputed database. This research presents findings, by compiling the most up-to-date state-of-the-art approaches using real-world data.