An Automatic Image Segmentation Technique Based on Pseudo-convex Hull

Abstract

This paper describes a novel method for image segmentation where imagecontains a dominant object. The method is applicable to a large classof images including noisy and poor quality images. It is fully automaticand has low computational cost. It may be noted that the proposedsegmentation technique may not produce optimal result in some casesbut it gives reasonably good result for almost all images of a largeclass. Hence, the method is found very useful for the applicationswhere accuracy of the segmentation is not very critical, e.g., forglobal shape feature extraction, second generation coding etc.