Research On Blind Identification Algorithm Of Image Splicing Based On Local Textural Features

With the rapid development of high-performance computer,a wide range of high-resolution image acquisition equipment and the use of powerful image processing software,digital images widely exist in our daily life.Because the image is intuitive and easy to understand,it is different from the text that the digital image can represent an effective way of communication between human beings.Therefore,the printed images in newspapers and video surveillance records provided as evidence in court are generally considered to be the proof of the authenticity of news reports and cases.The diversity and openness of social network channels make every person's life filled with a variety of digital images.While the extensive use of image editing software makes tampering with the contents of the image becomes very simple.Image tampering or forgery is no longer limited to experts or professional teams.A large number of image tampering events on the social media platform have greatly eroded the trust in the information of the visual content,which largely misled the audience.Therefore,in order to regain the trust of the image,it is urgent to develop a more reliable and effective digital image forensics technology to verify the authenticity of digital images.Digital image forensics technology has two main methods: digital image active forensics technology and digital image passive forensics technology.The active detection method tries to verify the authenticity of the image by detecting whether the digital watermark,signature or fingerprint is complete.However,the passive detection method is based on collecting through the evidence of image tampering,without any pre-embedded information.Due to the limitations of the active method,the passive detection method is more practical,and more and more researchers pay attention to it.As a basic and common passive image tampering method,digital image splicing has become one of the important research topics in digital image forensics.In this paper,we mainly focus on the study of image tampering detection methods.The background and significance of digital image forensics are introduced,and the active forensics technology and passive forensics technology of digital image are summarized respectively.Then,the principles and models of digital image splicing detection are briefly introduced and classified.The principle of the existing algorithm,the improvement of the algorithm,the experimental result of the algorithm and the limitation of the algorithm are studied respectively from the two main aspects of tampering traces and image content consistency.Finally,the image tampering detection algorithms based on statistical features and local texture features are introduced.These features are classified and summarized,and two improved algorithms are proposed based on local textural features.Texture,as the common but difficult to describe feature,the key to its description lies in the extraction of texture features.Aiming at the problem that the accuracy of the current image splicing detection algorithm is low and the local texture feature description is not sufficient,this paper presents a novel passive splicing detection method based on the multi-scale differential excitation information and Local tri-directional pattern(Ltridp)feature.Firstly,the color image to be detected is transformed into the YCb Cr color space,and the three-scale differential excitation features are extracted from on the chrominance channel(Cr and Cb)of the images.Then,the tri-directional pattern value magnitude pattern value of the original image are calculated,the tri-directional pattern value is converted into two binary patterns and the histogram of two binary patterns and the histogram of a magnitude pattern are calculate.Finally,the extracted histogram feature pairs are fused and sent to the support vector machine(SVM)classifier for training and detection,so as to judge the authenticity of the image.Experimental results show that the proposed algorithm has a significant improvement in detection rate,which has a certain significance in research.To further improve the detection rate with relatively low dimension feature vector,a novel passive splicing detection method using textural features based on the Gray Level Co-occurrence Matrices,namely TF-GLCM,is proposed in this paper.In the TF-GLCM,the GLCM are calculated based on the Difference Block Discrete Cosine Transform(DBDCT)arrays to capture the textural information and the spatial relationship between image pixels sufficiently.The discriminable properties contained in the GLCM are described by six textural features,which include two new introduced ones and four independent ones.In addition,the statistical moments mean Me and standard deviation SD of textural features are used instead of themselves as elements in feature vector to reduce the dimensionality of feature vector and computational complexity.A SVM is employed for classification purpose.Experimental results show that the TF-GLCM achieves the detection rates of 98% on CASIA v1.0,and 97% on CASIA v2.0 with 96-D feature vector.And the detection rates benefit from the two new textural features.Meanwhile,the TF-GLCM is superior to some state-of-the-art methods with lower dimension feature vector.