Research And Application Of Perceptual Hash Based On Human System

With the rapid development of computer network and multimedia technology, people can easily tamper and spread the digital images.The integrity of image content and authenticity of image information must be taken into account.Therefore the image authentication which can insure the integrity of image data and the authenticity of image content came into being.In recent years,the image authentication based on perceptual hashing has become a hot research topic.The perceptual information is extracted which will be compressed and coded to form the final hash sequence. Then the final hash sequences of the original image and the image to be certificated will be compared to judge the the image to be certificated is tempered or not.What image authentication essentially concern is the perceptual information of the image.Most of existing perceptual hashing algorithms extract image features without considering human visual system.A perceptual hashing algorithm based on human visual system is proposed.And based on this algorithm，a system is designed which can used to monitor the pirated images in Taobao.The main contributions and innovations are as follows:The basic structure and features of human visual system are analyzed and two classical models of human visual system are introduced.If the human visual model is applied to perceptual hashing algorithm,people’s subjective evaluation of the image will be in accord with the objective evaluation which direct the design of perceptual hashing algorithm proposed in this paper.The scheme proposed in chapter four is based on normalization and Watson Visual Model.The image is firstly normalized to make it robust against any affine transforms.With the Watson Visual Model,the measurement of perceptual hashing distance is consistent with people’s subject sense.Since a key is used in the algorithm, the fingerprint is hard to be forged.The experimental results indicate that the scheme is robust against modifications including geometric transformation such as rotation and scaling.Then,a pirated image monitoring system based on this algorithm is designed.The experiment results show that this system has high precision and recall.This system can used to monitor the pirated images in Taobao.