BRENDEL, Elliot (2016) Phase transition in Total Variation Regularization PRE - Research Project, ENSTA.



After formalising the phase transition issue for the Compressive Sensing (CS) and the Total Variation Regularization (TVR) cases, we attempt to compute bounds for the phase transition using a graph-theory point of view and adapting a scientific paper [1]. Second, we study the algorithmic aspect of the problem. We also try to compute directly the algorithmic phase transition and we expose our analysis of a modified algorithm. This work allows us to enlighten comparisons between algorithms. We finish being open on the question of the phase transition for practical applications.

Item Type:Thesis (PRE - Research Project)
Uncontrolled Keywords:Compressive sensing, total variation regularization, tomography, phase transition, GPU programming, reconstruction algorithms.
Subjects:Information and Communication Sciences and Technologies
Mathematics and Applications
Physics, Optics
ID Code:6732
Deposited By:Elliot Brendel
Deposited On:13 oct. 2016 10:48
Dernière modification:13 oct. 2016 10:48

Repository Staff Only: item control page