The Research On Doubly Compressed JPEF Image Passive Forcnsics

Image is one of the most intuitive ways of expression for information. Due tothe advantages of image which have large amount of information, it exists in a widerange of networks, newspapers and magazines. However, because the appearing ofthe various image processing algorithms, related image editing software andhardware, it easy to deal with images by non-professionals, which makes itimpossible to identify the image with naked eyes ,at the same time,this situationbring much trouble to our life, and even lead to a variety of crisis of confidence. Sothe detection of digital image authenticity is very important. This article studied onthe passive forensics of double compressed JPEG image, explained basic principlesof the JPEG image double compress operation, and used its internal features forimage detection. The main content as follows:Firstly, summaries theoretical background, architecture, classification andgeneral methods of digital image passive forensics, at the same time, introduce theresearch status at home and abroad.Secondly, analysis of the doubly compressed algorithms for JPEG images, thestatistical DCT coefficients characteristics of JPEG images has been obtained, andthe two methods common used for double JPEG compressed image detection arediscussed, which determined the thinking detection of this thesis.Thirdly, with the analysis of the internal quantization noise statisticalcharacteristics on the JPEG compressed image, the distribution characteristics areobtained. Then the detection whether the image has been tampered can be acquiredwith the distribution characteristics, and the tempering area can also be determined.Fourthly, because of the relationship of the pixel distribution after firstcompression or double compression in JPEG image, and analysis of the relationship.Then based on the relationship of the doubly compressed image, one method isproposed to detect the image tampering.At the end of this thesis, one summary is given out and the prospective developing of JPEG image passive forensics in the future is forecasted.