Structural and Textural Skeletons for Noisy Shapes

Abstract. The extraction of consistent skeletons in the presence of boundary noise is still a problem for most skeletonization algorithms. Many suppress skeletons associated with boundary perturbation, either by preventing their formation or removing them

Proceedings of the International Symposium on Visual Computing 2005 (ISVC2005), pp. 454-461

Structural and Textural Skeletons for Noisy Shapes

Wooi-Boon Goh and Kai-Yun Chan

School of Computer Engineering, Nanyang Technological University,

Nanyang Avenue, Singapore 639798 {aswbgoh, askychan}@ntu.edu.sg

Abstract. The extraction of consistent skeletons in the presence of boundary

noise is still a problem for most skeletonization algorithms. Many suppress

skeletons associated with boundary perturbation, either by preventing their

formation or removing them subsequently using additional operations. A more

appropriate approach is to view a shape as comprising of structural and textural

skeletons. The former describes the general structure of the shape and the latter

its boundary characteristics. These two types of skeletons should be encouraged

to remaining disconnected to facilitate gross shape matching without the need

for branch pruning. Such skeletons can be formed by means of a multi-

resolution gradient vector field (MGVF), which can be generated efficiently us-

ing a pyramidal framework. The robust scale-invariant extraction of the skele-

tons from the MGVF is described. Experimental results show that the MGVF

structural skeletons are less affected by boundary noise compared to skeletons

extract by other popular iterative and non-iterative techniques.

1 Introduction

Skeletons are widely adopted as a basis for shape description [4],[10]. Kimia [5] recently highlighted several psychophysical evidences that suggest some role for medial axis in the human visual system. However, it is still a challenge to extract robust skeletons in the presence of boundary noise. A shape’s skeleton can be defined as the locus of centers of maximal discs contained within the shape [2] and it is well-known that small boundary perturbations can significantly alter the structure of such skeletons. This problem has been traditionally addressed using some form of pre-smoothing technique or branch pruning strategy. In the case of the former, pre-smoothing the shape contour using curvature flow smoothing techniques [4] could be carried out before skeleton extraction. Alternatively, topology preserving pruning methods can be employed [11]. Multiscale pruning strategies such as the iteratively pruned trees of Voronoi edges have also been proposed [7]. Pizer et al. [8] described the medialness function (cores) of a shape over a width-proportional scale. By select-ing cores from appropriate scales, shapes can be made relatively insensitive to bound-ary noise. In summary, most approaches seek to suppress spurious skeletons either by preventing their formation or removing them subsequently. However, it has been acknowledged that selecting an appropriate scale or pruning threshold that can prop-

Word文档免费下载Word文档免费下载:Structural and Textural Skeletons for Noisy Shapes (共8页,当前第1页)

Structural and Textural Skeletons for Noisy Shapes相关文档

最新文档

返回顶部