A new approach to detect paste forgeries in an image

With the inception of advanced image altering and editing softwares such as Photoshop, GIMP and the like, the credibility of an image has always been a vital question at hand. Despite the ease with which an image has been tampered, we are still unable to visually distinguish between a forged and an authentic image, which calls for the requirement of a robust software to do the same. This paper presents a new algorithm to detect paste forgeries such as copy-move forgery and splicing in an image, using a bilateral filter and one level discrete wavelet transforms. A bilateral filter is a non-linear filter, that preserves edges while reducing noise and smoothing the image. This makes it easier to identify regions of sharp edges, which correspond to counterfeited regions in the image. This procedure showed good results when tried on images from an improved dataset that we created, consisting of spliced, copy move forged and original images.