Boundary Extraction in Images Using Hierarchical Clustering-based Segmentation

Hierarchical organization is one of the main characteristics of human segmentation. A human subject segments a natural image by identifying physical objects and marking their boundaries up to a certain level of detail [1]. Hierarchical clustering based segmentation (HCS) process mimics this capability of the human vision. The HCS process automatically generates a hierarchy of segmented images. The hierarchy represents the continuous merging of similar, spatially adjacent or disjoint, regions as the allowable threshold value of dissimilarity between regions, for merging, is gradually increased. HCS process is unsupervised and is completely data driven. This ensures that the segmentation process can be applied to any image, without any prior information about the image data and without any need for prior training of the segmentation process with the relevant image data.

The implementation details of HCS process have been described elsewhere in the author's work [2]. The purpose of the current study is to demonstrate the performance of the HCS process in outlining boundaries in images and its possible application in processing medical images.

[1] P. Arbelaez. Boundary Extraction in Natural Images Using Ultrametric Contour Maps. Proceedings 5th IEEE Workshop on Perceptual Organization in Computer Vision (POCV'06). June 2006. New York, USA.

[2] A. N. Selvan. Highlighting Dissimilarity in Medical Images Using Hierarchical Clustering Based Segmentation (HCS). M. Phil. dissertation, Faculty of Arts Computing Engineering and Sciences Sheffield Hallam Univ., Sheffield, UK, 2007.