The University of British Columbia

In this paper, we present a shape descriptor for content-based image retrieval, that is invariant to translation, rotation and scaling. The proposed shape descriptor, which is easy to compute, is based on a measure of how symmetric or asymmetric the shape of the object is. A discussion of how this shape descriptor can be used for content-based retrieval is provided. Moreover, experimental results that demonstrate the performance of the proposed shape descriptor on regular and irregular shapes are presented.