Title :
Identifiability and Noise Robustness for l1-Analysis Regularizations in Compressive Sensing
Author :
Nicod?me ;Turcu Flavius;Dossal Charles
Author_Institution :
Univ. de Bordeaux, Bordeaux, France
Abstract :
We use several geometric techniques to characterize identifiability and to estimate noise robustness in the framework of l1-analysis regularization. This extends several recent theoretical results and algorithms that deal with the same issues in the less complex case of l1-synthesis regularizations.
Keywords :
"Standards","Noise robustness","Noise measurement","Transforms","Algorithm design and analysis","Robustness","Compressed sensing"
Conference_Titel :
Symbolic and Numeric Algorithms for Scientific Computing (SYNASC), 2015 17th International Symposium on
DOI :
10.1109/SYNASC.2015.21
