Graph cuts code

This is code for a matlab implementation of graph cuts using L1 norm minimization.
The main file is best_gc.m and it requires usage of a conjugate gradient equation solver. So you can either use your own or use the version on this webpage, which is a suitably modified variant of Andreas Klimke's code available online. cg_klimke.m

The inputs to the graph cuts algorithm are a graph in the form of a node-edge adjacency matrix (A), an indicator vector (c), and a vector of weights on the edges (b). We worked on a simple binary labelling problem, namely foreground-background segmentation and a simplistic matlab program which creates graphs can be found here . It uses some gaussian mixture model functions from The GMMBAYES toolbox.