点击次数：

论文类型：期刊论文

发表时间：2013-01-01

发表刊物：Journal of Information and Computational Science

收录刊物：EI、Scopus

卷号：10

期号：13

页面范围：3981-3989

ISSN号：15487741

摘要：The local shape comparison is always necessary for many applications including visualization, shape matching, shape retrieval and so forth. In this paper a distance-based shape descriptor is proposed and used for local shape comparison. After regions to be compared are specified, the sampling points are extracted following the same rules from each region. Then, signed Euclidean distances from each sampling point of a region to an adjacent one are calculated and used to build the shape descriptor of that region. After that, the similarity of the given regions is evaluated by comparing their shape descriptors. The feasibility of the proposed method is proved by some experimental results. ? 2013 Binary Information Press.