Normalized Cuts and Image Segmentation

w e propose Q novel approach for solving the perceptual grouping problem in vision. Rather than focusing on local features and their consistencies in the amage data, our approach aims a t extracting the global impression of an image. We treat image segmentation QS (I graph partitioning problem and propose Q novel global criterion, the normalized cut, for segmenting the graph. The normalized cut craterion measures both the total dissimilarity between the different groups QS well as the total similarity within the groups. We show that an eficient computational technique based on a generaked eigenvalue problem can be used to optimize this criterion. w e have applied this approach to segmenting static images and found results very enco u raging.