摘要：The Fourier Descriptor (FD) is a powerful tool for shape analysis andmany signatures have been proposed to derive Fourier descriptors. These shapesignatures lack of important information in articulation and part structures ofcomplex shapes. In this study, the Inner-Centroid Distance (ICDs) signaturewhich is based on the Centroid Distance signature and Inner-Distance isdeveloped to overcome the shortcomings of existing signatures. The retrievalperformance is evaluated by using standard shape database and commonly usedperformance measurement. The experimental results demonstrate that the proposedsignature performs better than the comparison algorithm.
