A virtual instrument for efficient blind-source separation of nonstationary signals

In this paper two methods for blind source separation of nonstationary signals, such as electroencephalogram output, applied to time frequency distributions are compared through implementation in a virtual instrument. Both methods are based on image processing approaches, but adopt different strategies for solving the blind source separation problem: the first method is based on a data clustering extraction, while the second one relies on the initial estimation of number of components followed by an iterative peak detection and extraction algorithm. The proposed virtual instrument provides an efficient and fast method for medical signal analysis, with low execution time and low resource consumption.