A new technique for removing forensically detectable traces from digital images

The power of digital images on modern society is unbelievable. As the People become ever more depend on these images for visual communication. But storing images in a secure way has become very convoluted. There are a number of techniques that can be used to verify the authenticity of images. Despite the fact that many image forgery detection techniques are presented, the number of image tampering is increasing. As a result it is vital to find the limitations of offered detection methods to avoid further forgeries. In this paper, a new technique is considered to remove the forensically detectable traces from digital images. It also explains how it can be used to create unnoticeable forgeries. The proposed method works through the stimation, examination and alteration of the image´s transform coefficients. The existing tampering detection techniques make use of image compression history to find the forgeries. Here the proposed technique uses a specially designed noise to eliminate the indicators of compression. A proper addition of this specially designed noise helps the modified image to appear as an uncompressed one. This technique can be used in the similar way as existing methods to cover the history of operations performed on an image in the past in addition with that it can also be used to alter the digital signatures added to the images and there by rendering several forms of image tampering.