Modularity Based Color Image Segmentation

Abstract: Image segmentation directly influences the successive processing and analyzing of an image. Thus segmentation becomes one of the biggest problems in image processing. Inspired by the application of community detection algorithms in large-scale social networks, we attempt to view an image from the perspective of a network and consider the image segmentation problem as a community detection problem. For a network, modularity is a crucial quantity, which is used to evaluate the performance of various community detection algorithms. A graph based method with modularity, can avoid the over segmentation problem which is present in all traditional segmentation methods. Compared with other existing segmentation algorithms, proposed algorithms start with initial segmentation techniques and lead to the computation of modularity, and try to achieve better segmentation to some extent.

Keywords: Image Segmentation, Graph based image segmentation, Modularity.