Review of independent component analysis and its reconfigurable design

Independent component analysis (ICA) is essentially a method for extracting useful information from data. Independent component analysis finds underlying factors or components from multidimensional statistical data. ICA is distinguished from other methods in a way that, it looks for components that are both statistically independent, and non-gaussian. Since ICA algorithm is computationally complex and uses large volume of data sets, there is a need for technique that provides potentially faster and even real-time implementations for ICA algorithms for signal and image processing applications. Very large scale integration (VLSI) technology is a solution that provides Modularity, hierarchy, parallelism and satisfies these requirements. Reconfigurable modules play major role nowadays because they are highly reusable and ready to be retargeted to other ICA-related applications. However these solutions also have some limitations and Critical Challenges. This paper reviews basic concepts of ICA, existing methods of ICA, merits and demerits of its VLSI implementation. Though Review of ICA is done in several articles, review of ICA In VLSI, Major Challenges In ICA implementation is discussed in this paper in comprehensive manner.