A Blind Source Separation Algorithm Based on the KL Divergence and Feedback
Liu Zhongjia · Audio Engineering · 2015
By introducing the Kullback- Leibler divergence and feedback mechanism into nonnegative matrix factorization( NMF),a new blind source separation algorithm( KL- NMF) is presented. The KL divergence is used to measure the effects of nonnegative matrix factorization,and the correlation coefficient between the separated signal and the mixed signal is used to measure the purity of the separated signals. After each factorization,the most pure signal is extracted and a new mixed signal is obtained. Thus,all source signals are separated one by one. The simulation results show that the separation performance of the proposed algorithm is superior to the algorithm Euclidean distance based NMF.