0.5 Sparsity

[Christos Chatzichristos]

Goodevening to everybody,

I am trying to implement Thomas Schultz paper (1) for the auto-calibration of Spherical Deconvolution with sparsity constrain (with some adaptations). I know that using CVXOPT for optimisation step you can use sparsity =1 but for the 0,5 sparsity constrain that Thomas Schultz uses do you know how can i use CVXOPT with l0.5

Thank you in advance

(1) http://www.ncbi.nlm.nih.gov/pubmed/24505724