Radial Inverse Force Histograms: A new algorithm for content-based image retrieval

Abstract

This paper proposes a novel method, called "Radial Inverse Force Histograms (RIFH)". The principle of this method uses the gravitational forces in Newton's theory to represent the texture of image in histogram pattern. In a far and away distance the gravitational forces were reduced. In spite of this region has influence to retrieve the image. In order to overcome this problem, this research uses RIFH. The method can be used for retrieval invariant images (translation, rotation, and scaling). In experimental with the grey level image, this research studies and compare with the distance histograms method and angular histograms method. It found that, the RIFH has a better performance than other approaches. © 2010 IEEE.