期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2015
卷号：3
期号：4
DOI：10.15680/ijircce.2015.0304200
出版社：S&S Publications
摘要：Recently advanced image processing tools and computer graphics techniques make it straight forward toedit or modify digital images. In court, for police agencies, for insurance or media companies, this raises the challengeof discriminating original images from malicious forgeries. Particular region from an image is pasted into other imagewith purpose to create image splicing. Image splicing is a common type of image tampering (manipulation) operation. ,currently pictures can't be thought of trustworthy proof. In this project, machine learning approach is used that exploitsrefined inconsistencies within the color of the illumination of pictures, distinguishes between computer generated andphotographic image and exploits forged region in digital image. The technique is applicable to photographs containingtwo or additional individuals and normal non face images also and needs no knowledgeable interaction for thetampering decision. For differentiating the photorealistic computer generated image (PRCG) and photographic image(PIM), a completely unique approach is to concentrate on the image textures, and can acknowledge that pictures fromdigital cameras contain traces of resampling as a results of employing a color filter array with demosaicing parameters.
关键词：Digital tampering ; demosaicing; resampling; illuminant map; color inconsistencies; forgery
