A new Fourier shape descriptor using smallest rectangle distance

Abstract

Shape is one of the main features in content-based image retrieval (CBIR). This paper proposes a novel CBIR technique based on shape feature. This technique uses the distances between the boundary points of a shape and the smallest rectangle that covers it. The proposed technique is a Fourier based technique and it is invariant to translation, scaling and rotation. The retrieval performance between some commonly used Fourier based signatures and our Smallest Rectangle Distance (SRD) signature has been tested using MPEG-7 database. Experimental results are shown that the SRD signature has better performance compared with many of those signatures.