Abstract
Focal Underdetermined System Solver (FOCUSS) is a powerful method for sparse representation, in which the Lp-norm like cost function is very often used. However, this cost function is not only nondifferentiable but also can be very ill-conditioned in some situations. The local minima problem of FOCUSS is discussed in this paper. Moreover, to solve this problem, we first extend the Lp-norm like cost function to its corresponding Lp-approximation. After this, we analyze the nonconvexity of the new cost function, which results in that FOCUSS algorithm gets stuck in the local minima in many situations, especially when the hidden sources are not very sparse. To reduce the number of the local minima, a multilayer FOCUSS is developed in this paper. Comparing with the conventional FOCUSS, the experiments inclusing MRI reconstruction demonstrate that multilayer FOCUSS can significantly improve the performance. Even for some very challenging cases, where the conventional FOCUSS fails, multilayer FOCUSS still works well.
Similar content being viewed by others
References
He, Z.S., Xie, S.L., Fu, Y.L.: Sparsity analysis of signals. Prog. Nat. Sci. 16(8), 879–884 (2006)
He, Z.S., Xie, S.L., Fu, Y.L.: Sparse representation and blind source separation of Ill-posed mixtures. Sci. China F 49(5), 639–652 (2006)
He, Z.S., Cichocki, A., Zdunek, R., Cao, J.T.: CG-M-FOCUSS and Its Application to Distributed Compressed Sensing. Lecture Notes in Computer Science 5263, 237–245 (2008)
Xie, K., He, Z., Cichocki, A.: Convergence analysis of the FOCUSS algorithm. IEEE Trans. Neural Netw. Learn. Syst. 26(3), 601–613 (2015)
Xie, K., He, Z., Cichocki, A., Fang, X.: Rate of convergence of the FOCUSS algorithm. IEEE Trans. Neural Netw. Learn. Syst. doi:10.1109/TNNLS.2016.2532358
Aharon, M., Elad, M., Bruckstein, A.M.: The k-SVD: an algorithm for designing of overcomplete dictionaries for sparse representation. IEEE Trans. Signal Process. 54(11), 4311–4322 (2006)
Bofill, P., Zibulevsky, M.: Underdetermined blind source separation using sparse representations. Signal Process. 81, 2353–2362 (2001)
Chen, S., Donoho, D.L., Saunders, M.A.: Atomic decomposition by basis pursuit. SIAM J. Sci. Comput. 20(1), 33–61 (1998)
Cichocki, A., Amari, S.: Adaptive Blind Signal and Image Processing: Learning Algorithms and Applications. Wiley, New York (2003)
Cotter, S.F., Rao, B.D., Engan, K., Kreutz-Delgado, K.: Sparse solutions to linear inverse problems with multiple measurement vectors. IEEE Trans. Signal Process. 53(7), 2477–2488 (2005)
Donoho, D.L., Elad, M.: Maximal sparsity representation via \(L\)1 minimization. Proc. Natl. Acad. Sci. 100, 2197–2202 (2003)
Georgiev, P., Theis, F., Cichocki, A.: Sparse component analysis and blind source separation of underdetermined mixtures. IEEE Trans. Neural Netw. 16(4), 992–996 (2005)
Gorodnitsky, I.F., Rao, B.D.: Sparse signal reconstruction from limited data using FOCUSS: a recursive weighted minimum norm algorithm. IEEE Trans. Signal Process. 45(3), 600–616 (1997)
Kreutz-Delgado, K., Murry, J.F., Rao, B.D., et al.: Dictionary learning algorithms for sparse representation. Neural Comput. 15, 349–396 (2003)
Li, Y.Q., Amari, S., Cichocki, A., Ho, D.W.C., Xie, S.L.: Underdetermined blind source separation based on sparse representation. IEEE Trans. Signal Process. 54(2), 423–437 (2006)
Li, Y.Q., Cichocki, A., Amari, S.: Analysis of sparse representation and blind source separation. Neural Comput. 16, 1193–1234 (2004)
Olshausen, B.A., Field, D.J.: Emergence of simple-cell receptive field properties by learning a sparse code for natural images. Nature 381(6583), 607–609 (1996)
Rao, B.D.: Signal processing with the sparseness constraint. In: Proceedings of the ICASSP, vol. III, pp. 1861–1864. Seattle, WA (1998)
Rao, B.D., Gorodnitsky, I.F.: An affine scaling methodology for best basis selection. IEEE Trans. Signal Process. 47(1), 187–200 (1999)
Rao, B.D., Rngan, K., Cotter, S.F., Palmer, J., Kreutz-Delgado, K.: Subset selection in noise based on diversity measure minimization. IEEE Trans. Signal Process. 51(3), 760–770 (2003)
Saab, R., Yilmaz, O., Mckeown, M.J., Abugharbieh, R.: Underdetermined anechoic blind source separation via Lq-basis-pursuit. IEEE Trans. Signal Process. 55(8), 4004–4017 (2007)
Takigawa, I., Kudo, M., Toyama, J.: Performance analysis of minimum \(L\)1-norm solutions for underdetermined source separation. IEEE Trans. Signal Process. 52(3), 582–591 (2004)
Zhou, G., Cichocki, A., Zhang, Y., Mandic, D.: Group component analysis from multi-block data: common and individual feature extraction. IEEE Trans. Neural Netw. Learn. Syst. doi:10.1109/TNNLS.2015.2487364
Yang, Z., Xiang, Y., Xie, K., Lai, Y.: Adaptive method for nonsmooth nonnegative matrix factorization. IEEE Trans. Neural Netw. Learn. Syst. doi:10.1109/TNNLS.2016.2517096
Zhou, G., Zhao, Q., Zhang, Y., Adali, T., Xie, S., et al.: Linked component analysis from matrices to high-order tensors: applications to biomedical data. Proc. IEEE 104(2), 310–331 (2016)
Chen, C., Liu, Z., Zhang, Y., Philip Chen, C.L., Xie, S.:. Asymptotic fuzzy tracking control for a class of stochastic strict-feedback systems. IEEE Trans Fuzzy Syst. doi:10.1109/TFUZZ.2016.2566807
Chen, C., Liu, Z., Zhang, Y., Xie, S.: Coordinated motion/force control of humanoid robot with unknown sensor nonlinearity and manipulated object’s uncertainty. IEEE Trans. Syst. Man Cybern. doi:10.1109/TSMC.2016.2557219
Liu, Y., Pan, M., Yu, R., Zhang, Y., Xie, S.: SD-MAC: spectrum database-driven MAC protocol for cognitive machine-to-machine networks. IEEE Trans. Veh. Technol. doi:10.1109/TVT.2016.2555084
Liu, Y., Yuen, C., Yu, R., Zhang, Y., Xie, S.: Queuing-based energy consumption management for heterogeneous residential demands in smart grid. IEEE Transa. Smart Grid 7(3), 1650–1659 (2016)
Majumdar, Angshul: FOCUSS based Schatten-p norm minimization for real-time reconstruction of dynamic contrast enhanced MRI. IEEE Signal Process. Lett. 19(5), 315–318 (2012)
Kim, S.J., Koh, K., Lustig, M., Boyd, S., Gorinevsky, D.: An interiorpoint method for large-scale L1-regularized least squares. IEEE J. Sel. Topics Signal Process. 1(4), 606–617 (2007)
Gorodnitsky, I.F., George, J., Rao, B.D.: Neuromagnetic source imaging with FOCUSS: a recursive weighted minimum norm algorithm. Electroencephalogr. Clin. Neurophysiol. 95(4), 231–251 (1995)
Xu, P., Tian, Y., Chen, H.F., Yao, D.Z.: Lp norm iterative sparse solution for EEG source localization. IEEE Trans. Biomed. Eng. 54(3), 400–409 (2007)
Rao, B.D., Engan, K., Cotter, S.F., Palmer, J., Kreutz-Delgado, K.: Subset selection in noise based on diversity measure minimization. IEEE Trans. Signal Process. 51(3), 760–770 (2003)
Chen, J., Huo, X.M.: Theoretical results on sparse representations of multiple-measurement vectors. IEEE Trans. Signal Process. 54(12), 4634–4643 (2006)
He, Z.S., Xie, S.L., Ding, S.X., Cichocki, A.: Convolutive blind source separation in frequency domain based on sparse representation. IEEE Trans. Audio Speech Lang. Process. 15(5), 1551–1563 (2007)
He, Z.S., Xie, S.L., Zhang, L.Q., Cichocki, A.: A note on Lewicki–Sejnowski Gradient for learning overcomplete representation. Neural Comput. 20(3), 636–643 (2008)
He, Z.S., Cichocki, A., Zdunek, R., Xie, S.L.: Improved FOCUSS method with conjugate gradient (CG) iterations. IEEE Trans. Signal Process. 57(1), 399–404 (2009)
Acknowledgements
The authors would like to thank the National Natural Science Foundation of China (61322306, 61333013, 61603153, 61305036 and 61304140), Science and Technology Programs of Guangzhou of China (201508010007, 201604010007), the Natural Science Foundation of Guangdong Province under Grant S2011030002886 (team project), Scientific Funds approved in 2013 and 2016 for Higher Level Talents by Guangdong Provincial universities, Project supported by Guangdong Province Higher Vocational Colleges & Schools Pearl River Scholar Funded Scheme in 2014 for financial support. And, the second author’s research was also support by Guangdong Universities’ Technology Research Center for Satelliate Navigation Chips and applications.
Author information
Authors and Affiliations
Corresponding author
Appendix: Derivation of robust FOCUSS
Appendix: Derivation of robust FOCUSS
To solve optimization problem (6), here we similarly use the standard method of Lagrange multipliers to derive the iterative formula (7) for robust FOCUSS. Define the Lagrangian function \(L({\varvec{s}},{\varvec{\lambda }})\) as
A necessary condition for a minimizing solution sto exist is that (\({\varvec{s}}^{*},{\varvec{\lambda }}^{*})\) to be a stationary point of the Lagrangian function \(L({\varvec{s}},{\varvec{\lambda }})\). Thus, we have
where the partial derivative of \(J({\varvec{s}}^{*})\) with respect to \(s_i^*\) is
As in [5, 14], the expression (13) can be re-written as
where \({\varvec{\Pi }}({\varvec{s}}^{*})\!=\!diag\left\{ \left[ {\sqrt{(s_1^*)^{2}+\delta }} \right] ^{2-p},\left[ {\sqrt{(s_1^*)^{2}+\delta }} \right] ^{2-p}\!,\right. \cdots ,\left. \left[ {\sqrt{(s_n^*)^{2}+\delta }} \right] ^{2-p} \right\} \). From (12) and (14), we know that the stationary point (\({\varvec{s}}^{*},{\varvec{\lambda }}^{*})\) satisfies the following equation sets
From (15), carrying out some similar manipulations to [5, 14], we can derive
which leads to the following iterative formula for solving the optimization problem (6):
Rights and permissions
About this article
Cite this article
Xie, K., Shi, M., Wang, P. et al. A multilayer FOCUSS approach for sparse representation. Cluster Comput 20, 1325–1332 (2017). https://doi.org/10.1007/s10586-017-0823-6
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10586-017-0823-6