Random Walks Based Image Segmentation using Color Space Graphs

Image segmentation is one of the most involved topics of research in the area of Computer Vision and has attracted a lot of attention from the researchers. The goal of image segmentation is to change the representation of the image into something that is more meaningful and easier to analyze. In this process an image is partitioned into multiple segments by assigning label to pixels of the image so that the pixels sharing the same label can be characterized as a distinct object present in that image. Several automatic and semi-automatic approaches have been designed in order to segment an image. In this paper we will focus on a semi-automatic approach called the Random Walker approach which was first introduced by Leo Grady [1] and we will contribute several improvements which leads to better segmented images compared to the image segmentations achieved by the ideas involved in the above said paper. We will use graph theoretical approaches to partition the color space and give a new definition of the distance between two pixels that results in an improved segmentation. (C) 2013 The Authors. Published by Elsevier Ltd.