Subspace Pursuit for Compressive Sensing: Closing the Gap Between Performance and Complexity

By Wei Dai, Olgica Milenkovic. arXiv.org, 2008

1,698 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/b8cdd92fed11c8458785a91b45051b6e305e7b30

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Signal Recovery From Incomplete and Inaccurate Measurements Via Regularized Orthogonal Matching Pursuit (2007, methodology): “Compressive Sa mpling Matching Pursuit (CoSaMP) by Needell and Tropp and Subspace Pursuit (SP) by Dai and Milenkovic are greedy pursuits that incorporate ideas from ROMP, combinatorial algorithms, and convex optimization [19], [18], [10].”
Compressive Sensing in Electromagnetics - A Review (2015, methodology): “Otherwise, the retrieval of sparse scatterers located in the Fresnel region has been carried out in [24], where the CR formulation (15) of the arising CS problem has been addressed through the subspace pursuit approach in [81].”
ADMiRA: Atomic Decomposition for Minimum Rank Approximation (2009, methodology): “These include Compressive Sampling Matching Pursuit (CoSaMP) [13] and Subspace Pursuit (SP) [14].”
Generalized Orthogonal Matching Pursuit (2011, methodology): “In particul r, empirical frequency of exact reconstruction has been a popularly used tool to measure the ffectiveness of recovery algorithm [18], [25].”
Hard Thresholding Pursuit: An Algorithm for Compressive Sensing (2011, methodology): “The Compressive Sampling Matching Pursuit (CoSaMP) algorithm proposed by Needell and Tropp in [18] and the Subspace Pursuit (SP) algorithm proposed by Dai and Milenkovic in [8] do not provide better theoretical guarantees than the simple IHT algorithm, but they do offer better empirical performances”
Combining geometry and combinatorics: A unified approach to sparse signal recovery (2008, methodology): “Speciﬁcally, the papers [NV07, DM08, NT08] show that one can use greedy methods for data compressed using dense matrices satisfying the RIP property.”
CoSaMP: Iterative signal recovery from incomplete and inaccurate samples (2008, unknown): “See [10] for details.”
Normalized Iterative Hard Thresholding: Guaranteed Stability and Performance (2010, methodology): “For example, convex optimization can be used to find near-optimal estimates of [7] and similar bounds have been derived for estimates found by greedy algorithms such as Compressed Sensing Matching Pursuit (CoSaMP) [8], Subspace Pursuit [ 9 ], Iterative Hard Thresholding [10], and the method in [11].”
Sparse Representation for Target Detection in Hyperspectral Imagery (2011, methodology): “Alternatively, the problem in (11) can also be approximately solved by greedy pursuit algorithms such as orthogonal matching pursuit (OMP) [30] or subspace pursuit (SP) [31].”
Fast and Efficient Compressive Sensing Using Structurally Random Matrices (2011, background): “Furthermore, there also exists a family of greedy pursuit algorithms [6], [7], [8], [9], [10] offering another promising option for sparse reconstruction.”
Advances on Spectrum Sensing for Cognitive Radio Networks: Theory and Applications (2017, methodology): “These ca be listed as the regularized orthogonal matching pursuit (ROMP) [119], the compressive sampling matching pursuit (CoSaMP) [110], subspace pursuit (SP) [112], the tree-based orthogonal matching pursuit (TOMP) [111].”
Sparsity and Compressed Sensing in Radar Imaging (2010, background): “From the CS literature, a variant of orthogonal matching pursuit CoSaMP that iteratively selects and then prunes a group of coefficients [41], [42] provides stronger performance guarantees for stable signal recovery.”
Learning Sparsifying Transforms (2013, background): “Well-known examples of sparsifying transforms for natural signals and images include the Wavelets and DCT.”
Precise Undersampling Theorems (2010, unknown): “thresholding (red) [38], subspace pursuit (magenta) [37],”
Analysis of Orthogonal Matching Pursuit Using the Restricted Isometry Property (2009, methodology): “Moreover, the fact that our bound requires only the RIP of order K +1 is apparently unique among the published CS literature; much more common are results requiring the RIP of order 1.75K [12], 2K [11, 13], 3K [16, 18], 4K [15], and so on.”