Abstract
In this paper, we describe an implementation of the two-phase imagesegmentation algorithm proposed by Goldstein, Bresson, Osher in\cite{gold:bre}. This algorithm partitions the domain of a given 2d image intoforeground and background regions, and each pixel of the image is assignedmembership to one of these two regions. The underlying assumption for thesegmentation model is that the pixel values of the input image can besummarized by two distinct average values, and that the region boundaries aresmooth. Accordingly, the model is defined as an energy in which the variable isa region membership function to assign pixels to either region, originallyproposed by Chan and Vese in \cite{chan:vese}. This energy is the sum of imagedata terms in the regions and a length penalty for region boundaries.Goldstein, Bresson, Osher modify the energy of Chan-Vese in \cite{gold:bre} sothat their new energy can be minimized efficiently using the split Bregmanmethod to produce an equivalent two-phase segmentation. We provide a detailedimplementation of this method \cite{gold:bre}, and document its performancewith several images over a range of algorithm parameters.