Perceptual hashing for color image based on color opponent component and quadtree structure

Abstract

In this paper, a perceptual image hashing scheme based on color opponent component (COC) and Quadtree structure is proposed. Firstly, image normalization and Gaussian low-pass filtering are applied on input image to produce a secondary image, and then COC is extracted from the secondary image. The color change information is calculated from COC as color feature and Quadtree decomposition (QD) is applied on the intensity image of secondary image to produce Quadtree structure features. Finally, color features and Quadtree structure features are concatenated and pseudo-randomly permuted to produce final hash sequence. Experimental results show that the proposed scheme is robust to common content-preserving manipulations. The receiver operating characteristics (ROC) curves show that the proposed method outperforms some state-of-the-art methods in the performances of perceptual robustness and discrimination.