Blind Identification For Digital Images Based On Inconsistency Of Illuminant Direction

Image digitization is the pr oduct of rapid development of computermultimedia technology, and acquisition or printing equipments for digitalimages. It has been the sign of time transit ion to digital photos age. The rapiddevelopment and popularization of the image processing and edi ting software,so that non-professionals can edit digital images, modify, or even malicioustampering. Beautification of the digital image make people’s lives morecolorful, on the other hand, digital images are widely used in news reports,military, forensic, scientific research and medical fields. Digital imageauthenticity of identification technology to guarantee the authenticity of themedia images, judicial evidence image and military intelligence image as aneffective means to protect the security of digital image content, has importantand far-reaching significance.This article describes a common means of image tampering and the imageidentification techniques to identify the authenticity of digital image. Atpresent, there are three kinds of relati vely technology used to identify theauthenticity of digital images: Image identification technology based ondigital signature, Image identification technology based on digitalwatermarking, and Blind identification technology for image authenticity. Theembodiment of the image identification technology based on the digitalsignature and identification techniques are required based on the digital imagewatermarking method image provider prior to the image preprocessing; Blinddigital image identification t echnology does not require any pre-embeddedinformation under the premise of the identification of the authenticity of theimage, it has a wide range of application environments, and has graduallybecome The hot topic of scientific research.Image tampering will destroy the light consistency of natural images. Inorder to ensure the security of the image content, this paper presents a blindidentification method based on the consistency of illuminant direction to identify the image authenticity. In accordance with the error function between the measured intensity andcalculated intensity, and the regularization terms for different source models, this methodestimates the illuminant direction for images of different source models. Analyzing if theillumination directions between different objects and background in one image are consistentto determine whether the image has been tampered with. Experimental results showedthat the infinite light mode blind identification method of the digital image canbe effectively detected whether the image has been tampered with. In addition,experiments also show the effectiveness of blind identification base on local light source andcomposite light sources.