Integrating boundary cue with superpixel for image segmentation

This paper researches image segmentation as a global optimization problem and proposes a new way, which is called superpixel status model, to integrate boundary and region cue. Superpixel status model is a label model which describes the joint distribution of boundary and region classification in a bayesian framework. For organizing a boundary classifier, the contour of superpixel is decomposed into multiple line segments, and a robust line descriptor is presented to form line feature vector. Finally, an objective function is defined to assemble all superpixels statuses across the entire image for segmentation. Experiments and results show that the effectiveness of our approach.