Author : Shweta P. Kachhawal and Avinash P. Wadhe

Nowadays there’s great accessibility of tools for the acquisition and process of transmission signals. With this refined and cheap image editing software, it’s changing into more easier task to tamper with digital pictures. Therefore, currently pictures can’t be thought of trustworthy proof. To cope with these problems, understanding the sort of image manipulation is vital. This paper presents an summary of machine learning approach that exploits refined inconsistencies within the color of the illumination of pictures and distinguishes between computer generated and photographic image. The technique is applicable to photographs containing two or additional individuals and needs no knowledgeable interaction for the tampering decision. To achieve this, the data is taken from physics and statistical based illuminant estimator on image regions of comparable material. For differentiating the PRCG and PIM, a completely unique approach is to concentrate on the image textures, we have a tendency to acknowledge that pictures from digital cameras that contain traces of resampling as a results of employing a color filter array with demosaicing algorithms.

Keywords: Digital tampering, demosaicing, resampling, illuminant map, color inconsistencies, Digital forgery

Article published in International Journal of Current Engineering and Technology, Vol.5, No.2 (April-2015)