What is a detailed explanation of the procedure to learn the parameters w for a CRF given the energy function ?

This paper proposes an energy function and cross-validation, suitable for my purpose: superpixel semantic segmentation, but I do not understand what to do to learn parameters w. At this stage I am able to represent each superpixel by a histogram and consequently by an SVM probability estimate.

asked Sep 08 '14 at 07:00

Valentin%20Gheorghe's gravatar image

Valentin Gheorghe
1112

edited Sep 08 '14 at 08:04

Be the first one to answer this question!
toggle preview

powered by OSQA

User submitted content is under Creative Commons: Attribution - Share Alike; Other things copyright (C) 2010, MetaOptimize LLC.