Underdetermined Mixture Matrix Estimation Based on Neural Network and Genetic Algorithm
This paper proposes an improved approach to estimate the underdetermined mixture matrix to improve the performance of underdetermined blind source separation (UBSS) for speech sources. This approach only use two observed signals and consider a tangent value instead of each vector of the mixture matrix for estimation. An improved clustering method based on competitive neural network and genetic algorithm is then designed to estimate these tangent values. In the proposed method, those tangent values are designed as clustering centers. The competitive neural network is used first to obtain the initial clustering centers, and genetic algorithm is applied to search for the global optimum around the initial clustering centers. Experimental results show that the tangent values of the observed vectors have better clustering characteristics, which could reduce the computational complexity for mixture matrix estimation. The improved clustering algorithm based on neural network and genetic algorithm can estimate a better mixture matrix with high precision than the general neural network clustering algorithm, and it can improve the performance of underdetermined blind signal separation.
KeywordsUnderdetermined blind source separation Tangent value Competitive neural network Genetic algorithm
This work was supported by the National Natural Science Foundation of China (Grant No. 61401145), the Natural Science Foundation of Jiangsu Province (Grant No. BK20140858, BK20151501), and the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).
- 3.Zhen, L., Peng, D., Yi, Z., Xiang, Y., Chen, P.: Underdetermined blind source separation using sparse coding. IEEE Trans. Neural Netw. Learn. Syst. PP(99), 1–7 (2015)Google Scholar
- 4.Li, Y., Cichocki, A., Amari, S.I.: Sparse component analysis for blind source separation with less sensors than sources. In: Independent Component Analysis (2010)Google Scholar
- 6.Cichock, A., Kasprzak, W., Amari, S.I.: Neural network approach to blind separation and enhancement of images. In: 1996 8th European Signal Processing Conference (EUSIPCO 1996), Trieste, Italy, pp. 1–4 (1996)Google Scholar
- 7.Pradhan, D., Wang, S., Ali, S., Yue, T., Liaaen, M.: CBGA-ES: a cluster-based genetic algorithm with elitist selection for supporting multi-objective test optimization. In: 2017 IEEE International Conference on Software Testing, Verification and Validation (ICST), Tokyo, Japan, pp. 367–378 (2017)Google Scholar
- 9.Zhang, M., Li, X., Peng, J.: Blind source separation using joint canonical decomposition of two higher order tensors. In: 2017 51st Annual Conference on Information Sciences and Systems (CISS), Baltimore, MD, pp. 1–6 (2017)Google Scholar
- 10.Bofill, P., Zibulevsky, M.: Sound examples. http://www.ac.upc.es/homes/pau/