Labeling Still Image Databases Using Graphical Models
G JoseI · 2009
Graphical models have proved to be very efficient models for labeling image data. In this paper, the use of graphical models based on Decomposable Triangulated Graphs are applied for several still image databases landmark localization. We use a recently presented algorithm based on the Branch&Bound methodology, that is able to improve the state of the art. Experimental results show the improvement given by this new algorithm with respect to the classical Dynamic Programming based approach.