PCA based shape descriptors for shape retrieval and the evaluations

Shape information is an important distribution to Content-Base Image Retrieval systems. In this paper we present a Principal Components Descriptor based on principal component analysis, as well as an enhanced Principal Components Descriptor for shape-based image retrieval. Experiments are performed on the MPEG-7 CE1-B test database. The presented descriptors are statistically evaluated by comparing their performance with several major shape descriptors, including the Fourier Descriptor the Wavelet Descriptor and the Curvature Scale Space Descriptor The evaluation reveals that, among all the five shape descriptors, EPCD presents the highest robustness to various shape poses, and is more efficient than other descriptors.