Abstract
This paper presents a framework for compressed sensing that bridges a gap between existing theory and the current use of compressed sensing in many realworld applications. In doing so, it also introduces a new sampling method that yields substantially improved recovery over existing techniques. In many applications of compressed sensing, including medical imaging, the standard principles of incoherence and sparsity are lacking. Whilst compressed sensing is often used successfully in such applications, it is done largely without mathematical explanation. The framework introduced in this paper provides such a justification. It does so by replacing these standard principles with three more general concepts: asymptotic sparsity, asymptotic incoherence and multilevel random subsampling. Moreover, not only does this work provide such a theoretical justification, it explains several key phenomena witnessed in practice. In particular, and unlike the standard theory, this work demonstrates the dependence of optimal sampling strategies on both the incoherence structure of the sampling operator and on the structure of the signal to be recovered. Another key consequence of this framework is the introduction of a new structured sampling method that exploits these phenomena to achieve significant improvements over current stateoftheart techniques.
Original language  English 

Article number  e4 
Number of pages  84 
Journal  Forum of Mathematics, Sigma 
Volume  5 
Early online date  15 Feb 2017 
DOIs  
Publication status  Published  28 Feb 2017 
Fingerprint Dive into the research topics of 'Breaking the coherence barrier: A new theory for compressed sensing'. Together they form a unique fingerprint.
Profiles

Clarice Poon
Person: Research & Teaching