Research On Digital Image Tamper Blind Detection

Digital image forensics technology is mainly divided into two categories: active forensics and passive forensics.Active forensics refers to adding a watermark or signature into images when we get them.If we want to detect an image which has been tampered,we only need to detect the integrity of the watermark.This method has certain limitations.Passive forensics is also known as blind forensics,premising that does not rely on any pre-embedded information,according to the property of the images original,the authenticity and integrity of the image detection,so the digital forensics is the main direction of the current digital image forensics technology,which has important application value.In this paper,two kinds of commonly used methods of tampering-splicing tampering and copy-paste tampering are studied.The main research work of this paper is as follows:Firstly,the basic concepts of digital image forensics technology are introduced,and the research status at home and abroad is analyzed.This paper introduces the blind forensics of digital image,and describes the classification of image blind forensics.This paper summarizes the two main methods of tampering,enumerates some existing algorithms,analyzes the advantages and disadvantages of these algorithms,and provides theoretical supports for the research.Secondly,a digital image splicing blind detection algorithm based on local mean decomposition(LMD)and moment features is proposed.The image detection problem can be translated into two classification problems in pattern recognition.We extracted the LMD(local mean decomposition)features and moment features from the image respectively,and we get the mixed features.Then we choose Adaboost algorithm to classify.The experimental results show that the algorithm has high detection accuracy and low computational complexity,and are superior to the other commonly used detection algorithms in Columbia University digital image splicing detection database.Finally,a blind detection algorithm based on complex moments is proposed for copy paste operation.In this method,discrete wavelet transform is used to reduce the dimension of the image.And we divided the image into the same size overlap blocks.Then we extracted complex moment features of each image block.Then,each image block is matched to detect and locate the copy paste region.The experimental results showed that the algorithm had higher detection accuracy and detection quality compared with the existing algorithms.