A Perceptual Hashing Algorithm Based on Adjustable Visual Threshold

Abstract

The researches on Human Visual System (HVS) lead to more comprehension in the visual perceptibility of human eyes and make dramatic contribution in image quality assessment. However, few perceptual hashing algorithms take the visual perceptibility into account. In this paper, we propose a new visual threshold adjustable perceptual hashing method under the guidance of Watson's visual model. This method encodes the perception information in the phase of feature extraction and coding, and uses a visual threshold parameter to reflect the user's acceptable image distortion range. Normally, different application declare different visual threshold to decide the acceptable visual error, our method can improve the discriminability near the given point of application according to the visual threshold required.