TY - JOUR
T1 - A greedy algorithm for the analysis transform domain
AU - Giryes, Raja
N1 - Publisher Copyright:
© 2015 Elsevier B.V.
PY - 2016/1/15
Y1 - 2016/1/15
N2 - Many image processing applications benefited remarkably from the theory of sparsity. One model of sparsity is the cosparse analysis one. It was shown that using ℓ1-minimization one might stably recover a cosparse signal from a small set of random linear measurements if the operator is a frame. Another effort has provided guarantee for dictionaries that have a near optimal projection procedure using greedy-like algorithms. However, no claims have been given for frames. A common drawback of all these existing techniques is their high computational cost for large dimensional problems. In this work we propose a new greedy-like technique with theoretical recovery guarantees for frames as the analysis operator, closing the gap between greedy and relaxation techniques. Our results cover both the case of bounded adversarial noise, where we show that the algorithm provides us with a stable reconstruction, and the one of random Gaussian noise, for which we prove that it has a denoising effect, closing another gap in the analysis framework. Our proposed program, unlike the previous greedy-like ones that solely act in the signal domain, operates mainly in the analysis operator[U+05F3]s transform domain. Besides the theoretical benefit, the main advantage of this strategy is its computational efficiency that makes it easily applicable to visually big data. We demonstrate its performance on several high dimensional images.
AB - Many image processing applications benefited remarkably from the theory of sparsity. One model of sparsity is the cosparse analysis one. It was shown that using ℓ1-minimization one might stably recover a cosparse signal from a small set of random linear measurements if the operator is a frame. Another effort has provided guarantee for dictionaries that have a near optimal projection procedure using greedy-like algorithms. However, no claims have been given for frames. A common drawback of all these existing techniques is their high computational cost for large dimensional problems. In this work we propose a new greedy-like technique with theoretical recovery guarantees for frames as the analysis operator, closing the gap between greedy and relaxation techniques. Our results cover both the case of bounded adversarial noise, where we show that the algorithm provides us with a stable reconstruction, and the one of random Gaussian noise, for which we prove that it has a denoising effect, closing another gap in the analysis framework. Our proposed program, unlike the previous greedy-like ones that solely act in the signal domain, operates mainly in the analysis operator[U+05F3]s transform domain. Besides the theoretical benefit, the main advantage of this strategy is its computational efficiency that makes it easily applicable to visually big data. We demonstrate its performance on several high dimensional images.
KW - Analysis
KW - Compressed sensing
KW - Greedy algorithms
KW - Iterative hard thresholding
KW - Sparse representations
KW - Sparsity for big data
UR - http://www.scopus.com/inward/record.url?scp=84948718411&partnerID=8YFLogxK
U2 - 10.1016/j.neucom.2015.02.100
DO - 10.1016/j.neucom.2015.02.100
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:84948718411
SN - 0925-2312
VL - 173
SP - 278
EP - 289
JO - Neurocomputing
JF - Neurocomputing
ER -