STUDYING THE FEASIBILITY AND IMPORTANCE OF GRAPH-BASED IMAGE SEGMENTATION TECHNIQUES
S. V. Kasmir Raja, Abdelkrim Khadir, S. S. Riaz Ahamed · 2009
Image segmentation and its performance evaluation are very difficult but important problems in computer vision. A major challenge in segmentation evaluation comes from the fundamental conflict between generality and objectivity: For general-purpose segmentation, the ground truth and segmentation accuracy may not be well defined, while embedding the evaluation in a specific application, the evaluation results may not be extensible to other applications. This paper analyzes the performance of Normalized Cut (NC) and Efficient Graph (EG) methods of Image Segmentation. We treat image segmentation as graph partitioning problem and propose novel global criterion, NC for segmenting the graph. The NC criterion measures both total dissimilarity between the different groups as well as the total similarity within the groups. We apply efficient graph based image segmentation method to image segmentation using two different kinds of local neighbourhood in constructing the graph. We also present a special strategy to compare and analysis the above two graph-based methods