Image Authentication by Detecting Traces of Demosaicing and Color Classification Methods
Recently advanced image processing tools and computer graphics techniques make it straight forward to edit or modify digital images. In court, for police agencies, for insurance or media companies, this raises the challenge of discriminating original images from malicious forgeries. Particular region from an image is pasted into other image with 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 exploits refined inconsistencies within the color of the illumination of pictures, distinguishes between computer generated and photographic image and exploits forged region in digital image. The technique is applicable to photographs containing two or additional individuals and normal non face images also and needs no knowledgeable interaction for the tampering 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 from digital cameras contain traces of resampling as a results of employing a color filter array with demosaicing parameters.