Abstract
As an important factor in the fast fixed-point algorithm of independent component analysis (ICA), noise has a significant influence on the separate performance of ICA. Unfortunately, the traditional algorithm of noisy ICA did not address the influence of impulsive components. Because the sources were signals mixed with impulsive noise, the Gaussian noisy algorithm will be invalid for separating the sources. In general, those measurements that significantly deviate from the normal pattern of sensed data are considered impulses. In this paper, we introduce a non-linear function based on the S-estimator to identify the impulsive components in the observed data. This approach guarantees that the impulse noise can be detected from the observed signal. Furthermore, a threshold for the impulse components and methods to remove impulse noise and reconstruct the signal is proposed. The proposed technique improves the separate performance of the traditional algorithm for Gaussian noisy ICA. With the proposed method, the fast fixed-point algorithm of ICA is more reliable for noisy situations. The simulation results show the effectiveness of the proposed method.
Similar content being viewed by others
References
Liu, K., Tian, X., & Cai, L. (2015). A noisy independent component analysis algorithm with low signal-to-noise ratio. Control Engineering of China, 2, 027.
Cai, L., & Tian, X. (2015). A new process monitoring method based on noisy time structure independent component analysis. Chinese Journal of Chemical Engineering, 23(1), 162–172.
Nassiri, V., Aminghafari, M., & Mohammad-Djafari, A. (2014). Solving noisy ICA using multivariate wavelet denoising with an application to noisy latent variables regression. Communications in Statistics-Theory and Methods, 43(10–12), 2297–2310.
He, X., & Zhu, T. (2014). ICA of noisy music audio mixtures based on iterative shrinkage denoising and FastICA using rational nonlinearities. Circuits, Systems, and Signal Processing, 33(6), 1917–1956.
Hyvärinen, A. (1999). Fast and robust fixed-point algorithms for independent component analysis. IEEE Transactions on Neural Networks, 10(3), 626–634.
Hyvärinen, A., & Oja, E. (2000). Independent component analysis: Algorithms and applications. Neural networks, 13(4), 411–430.
Hyvärinen, A., & Oja, E. (1997). A fast fixed-point algorithm for independent component analysis. Neural Computation, 9(7), 1483–1492.
Hyvärinen, A., Karhunen, J., & Oja, E. (2004). Independent component analysis. London: Wiley.
Bell, A. J. (2000). Information theory, independent-component analysis, and applications. Unsupervised Adaptive Filtering, 1, 237–264.
Bell, A. J., & Sejnowski, T. J. (1995). An information–maximization approach to blind separation and blind deconvolution. Neural Computation, 7(6), 1129–1159.
Gaeta, M., & Lacoume, J. L. (1990). Source separation without a priori knowledge: The maximum likelihood solution. Proceedings of EUSIPCO, 90, 621–624.
Pham, D. T. (1996). Blind separation of instantaneous mixture of sources via an independent component analysis. IEEE Transactions on Signal Processing, 44(11), 2768–2779.
Pham, D. T., & Garat, P. (1997). Blind separation of mixture of independent sources through a quasi-maximum likelihood approach. IEEE Transactions on Signal Processing, 45(7), 1712–1725.
Pahm, D. T, Garrat, P., & Jutten, C. (1992). Separation of a mixture of independent sources through a ML approach. In Proceedings of European signal processing conference (p. 771).
Moulines, E., Cardoso, J. F., & Gassiat, E. (1997). Maximum likelihood for blind separation and deconvolution of noisy signals using mixture models. In IEEE international conference on acoustics, speech, and signal processing, 1997. ICASSP-97 (Vol. 5, pp. 3617–3620). IEEE.
Pajunen, P., & Karhunen, J. (1997). Least-squares methods for blind source separation based on nonlinear PCA. International Journal of Neural Systems, 8(05–06), 601–612.
Hyvärinen, A. (1999). Fast ICA for noisy data using Gaussian moments. In Proceedings of IEEE international symposium on circuits and systems, 1999. ISCAS’99 (Vol. 5, pp. 57–61). IEEE.
Hyvärinen, A. (1999). Gaussian moments for noisy independent component analysis. IEEE Signal Processing Letters, 6(6), 145–147.
Koivunen, V., Enescu, M., & Oja, E. (2001). Adaptive algorithm for blind separation from noisy time-varying mixtures. Neural Computation, 13(10), 2339–2357.
Koivunen, V., & Oja, E. (1999). Predictor–corrector structure for real-time blind separation from noisy mixtures. ICA, 99, 479–484.
Hyvärinen, A. (1999). Sparse code shrinkage: Denoising of nongaussian data by maximum likelihood estimation. Neural Computation, 11(7), 1739–1768.
Hyvärinen, A., Hoyer, P., & Oja, E. (1999). Image denoising by sparse code shrinkage. In Intelligent signal processing. IEEE Press.
Rousseeuw, P. J., & Leroy, A. M. (2005). Robust regression and outlier detection. London: Wiley.
Sengijpta, S. K. (1995). Fundamentals of statistical signal processing: Estimation theory. Technometrics, 37(4), 465–466.
Therrien, C. W. (1992). Discrete random signals and statistical signal processing. New Jersey: Prentice Hall PTR.
Amari, S., Cichocki, A. &, Yang, H. H. (1996). A new learning algorithm for blind signal separation. In D. Touretzky, M. Mozer, M. Hasselmo (Eds.), Advances in neural information processing systems.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
1.1 Proof Lemma 1
We first suppose the case l odd (l = 2q + 1). Using S = s(r 1, …, r l ) for ease of notation, we can obtain
Suppose med i |r i | > cS. Because \(med_{i} \left| {r_{i} } \right| = \left| r \right|_{{q + 1{\kern 1pt} :l}}\), it holds that at least q + 1 of the \(\frac{{\left| {r_{i} } \right|}}{S}\) is larger than c. Consequently,
Therefore, \(med_{i} \left| {r_{i} } \right| \le cS\).
Now, we suppose that \(\text{g}^{ - 1} \left( {\frac{{\text{g} \left( c \right)}}{l + 1}} \right)S > med_{i} \left| {r_{i} } \right|\). This would imply that the first q + 1 of the \(\frac{{\left| {r_{i} } \right|}}{S}\) is strictly smaller than \(\text{g}^{ - 1} \left( {\frac{{\text{g} \left( c \right)}}{l + 1}} \right)\). Introducing this in \(\frac{1}{l}\sum\nolimits_{i = 1}^{l} {{\text{g}}\left( {\frac{{\left| {r_{i} } \right|}}{S}} \right)}\), we find that
Therefore, \(\text{g}^{ - 1} \left( {\frac{g\left( c \right)}{l + 1}} \right)S \le med_{i} \left| {r_{i} } \right|\).
Now, n is even (l = 2q), which proves that
Because \(med_{i} \left| {r_{i} } \right| = \frac{1}{2}\left( {\left| r \right|_{{q{\kern 1pt} :l}} + \left| r \right|_{q + 1:l} } \right)\), we have that at least q of the \(\frac{{\left| {r_{i} } \right|}}{s}\) is strictly larger than c, and
Except when all other |r i | are zero, the set of solutions for Eq. (10) is the interval \(\left( {0,\frac{{2\,med_{i} \left| {r_{i} } \right|}}{c}} \right]\). Therefore, \(S = \frac{{2\,med_{i} \left| {r_{i} } \right|}}{c}\). In other cases, med i |r i | ≤ cS.
Suppose now \(med_{i} \left| {r_{i} } \right| < \frac{1}{2}\text{g}^{ - 1} \left( {\frac{{2\text{g} \left( c \right)}}{l + 2}} \right)S\). Then, \(\left| r \right|_{{q + 1{\kern 1pt} {\kern 1pt} :{\kern 1pt} {\kern 1pt} n}} < \text{g}^{ - 1} \left( {\frac{{2\text{g} \left( c \right)}}{l + 2}} \right)S\). Hence, the first q + 1 of the \(\frac{{\left| {r_{i} } \right|}}{S}\) is less than \(g^{ - 1} \left( {\frac{{2\text{g} \left( c \right)}}{l + 2}} \right)\). Then, we obtain
That means \(med_{i} \left| {r_{i} } \right| \ge \frac{1}{2}\text{g}^{ - 1} \left( {\frac{{2\text{g} \left( c \right)}}{l + 2}} \right)S\).
Rights and permissions
About this article
Cite this article
Feng, P., Li, L. On extending the Noisy Independent Component Analysis to Impulsive Components. Wireless Pers Commun 88, 415–427 (2016). https://doi.org/10.1007/s11277-015-3135-2
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11277-015-3135-2