Signal separation method using ICA

Recently, multimedia systems have required more efficient signal separation methods to preserve the quality of voice or music recording in a noisy environment. Some of signal separation methods are based on minimizing the dependent measure between input signals to separate the noise component since this noise component is usually independent of the other signals. Under such circumstances, we have developed a new method to separate independent signal components which directly minimizes the Kullback-Leibler divergence using a genetic algorithm. We propose an improved method using differential information. As the result of the simulation, the separated signals are clearly separated into the original signals