A Genetic Approach to Fusion of Algorithms for Compressive Sensing

Inspired by the data fusion principle, we proposed a genetic approach to fusion of algorithms for CS to improve reconstruction performance. Firstly, several compressive sensing reconstruction algorithms (CSRAs) are executed in parallel to provide their estimates of the underlying sparse signal. Next, genetic algorithm is used to fuse these estimates for achieving a new estimate that is better than the best of these estimates. The proposed approach provides flexible design of fitness function and mutation strategy of genetic algorithm, and various participating CSRAs can be used to recover the sparse signal. Experiments were conducted on both synthetic and real world signals. Results indicate that the proposed approach has three advantages: (1) it performs well even when the dimension of measurements is very low, (2) reconstruction performance is better than any participating CSRAs, and (3) it is comparable or even superior to other fusion algorithm like FACS.

[1]  Lawrence Carin,et al.  Bayesian Compressive Sensing , 2008, IEEE Transactions on Signal Processing.

[2]  David L Donoho,et al.  Compressed sensing , 2006, IEEE Transactions on Information Theory.

[3]  J. Tropp,et al.  CoSaMP: Iterative signal recovery from incomplete and inaccurate samples , 2008, Commun. ACM.

[4]  Joel A. Tropp,et al.  Signal Recovery From Random Measurements Via Orthogonal Matching Pursuit , 2007, IEEE Transactions on Information Theory.

[5]  K. V. S. Hari,et al.  Progressive fusion of reconstruction algorithms for low latency applications in compressed sensing , 2014, Signal Process..

[6]  E.J. Candes,et al.  An Introduction To Compressive Sampling , 2008, IEEE Signal Processing Magazine.

[7]  K. V. S. Hari,et al.  Fusion of Algorithms for Compressed Sensing , 2013, IEEE Trans. Signal Process..

[8]  Aswin C. Sankaranarayanan,et al.  Compressive Sensing , 2008, Computer Vision, A Reference Guide.

[9]  Lothar M. Schmitt,et al.  Asymptotic Convergence of Scaled Genetic Algorithms to Global Optima , 2004 .

[10]  G.B. Moody,et al.  The impact of the MIT-BIH Arrhythmia Database , 2001, IEEE Engineering in Medicine and Biology Magazine.

[11]  Wei Lu,et al.  Regularized Modified BPDN for Noisy Sparse Reconstruction With Partial Erroneous Support and Signal Value Knowledge , 2010, IEEE Transactions on Signal Processing.

[12]  Mohamed Abdel-Mottaleb,et al.  Discriminant Correlation Analysis: Real-Time Feature Level Fusion for Multimodal Biometric Recognition , 2016, IEEE Transactions on Information Forensics and Security.

[13]  Volkan Cevher,et al.  Model-Based Compressive Sensing , 2008, IEEE Transactions on Information Theory.

[14]  Olgica Milenkovic,et al.  Subspace Pursuit for Compressive Sensing Signal Reconstruction , 2008, IEEE Transactions on Information Theory.

[15]  Jean-Luc Starck,et al.  Very high quality image restoration by combining wavelets and curvelets , 2001, SPIE Optics + Photonics.

[16]  Michael Elad,et al.  A Plurality of Sparse Representations Is Better Than the Sparsest One Alone , 2009, IEEE Transactions on Information Theory.