Abstract #1804

Parameter selection in Total Generalized Variation based reconstruction problems.

Carlos Milovic^1,2, Jose Miguel Pinto^1,2, Julio Acosta-Cabronero³, Pablo Irarrazaval^1,2, and Cristian Tejos^1,2

¹Department of Electrical Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile, ²Biomedical Imaging Center, Pontificia Universidad Catolica de Chile, Santiago, Chile, ³German Center for Neurodegenerative Deceases (DZNE), Magdeburg, Germany

An strategy for parameter selection in TGV regularized reconstruction problems is presented, with applications to deconvolutions and QSM. This allows fine-tuning of parameters in an efficient way and the use of predictors that are correlated to optimized results in terms of MSRE. This allows users to automatize or accelerate the parameter selection, critical in expensive problems such as QSM and reduce the error in the reconstruction.

This abstract and the presentation materials are available to members only; a login is required.

Join Here