Image Hashing Using Ring Partition and Invariant Vector Distance

Abstract

As well famed, In image hashing there are two important objectives robustness and discrimination. With the help of ring partition and invariant vector distance we can improve the robustness and discriminative capability if image hash algorithm. As ring partition not change the angle of any images i.e. image rotation and statistical features are combined together from different rings of images to the perceptually uniform color space. Euclidean distance between the vectors of the featured images are not changed on different operations (Brightness, rotation, noise, blur, contrast, etc) which help of image hash to make the image more compact and discriminative. We have conduct an experiment on dataset i.e, Barcelona to evaluate the efficiency and to demonstrate the hashing algorithm to different operations performed on images.