Image forgery detection based on illumination inconsistencies amp; intrinsic resampling properties

摘要

Photographs are used to represent real-world events. Today many powerful image editing software's like Gimp, Photoshop etc. are available. Any kind of images can be manipulated by using these software's. Image composition or splicing is one of the important image manipulation technique. Sometimes these manipulated images are provided as evidence in court and this may cause serious problems. So it is important to check whether the images available are forged or not. So based on the challenges of forgery detection here a new method is proposed which incorporates both the illumination inconsistencies and resampling properties for detecting forged images. Illumination inconsistency is due to the fact that while creating a forged image it is difficult to achieve proper illuminant condition for the entire image. Also while creating a forged image it is often necessary to resize certain portions of image and this requires resampling image into a new sampling lattice. This introduces certain resampling properties which can be measured. This technique can be used in medical as well as forensic applications to check the genuinity of images.