Detection of Region Duplication Forgery in Digital Images Using Wavelets and Log-Polar Mapping

Abstract

Due to the abundantly available imaging technologies, manipulation of digital images has become a serious problem nowadays, in various fields like medical imaging, digital forensics, journalism, scientific publications, etc. In this paper, we concentrate on detection of a specific category of digital image forgery known as region duplication forgery or copy-move forgery, which is done by copying a block of an image and pasting it on to some other block of the same image. We present a novel approach based on the application of wavelet transform that detects and localizes such forgeries. Our technique works by first applying wavelet transform to the input image to yield a reduced dimension representation. We then perform exhaustive search to identify the similar blocks in the image by mapping them to log-polar coordinates and using phase correlation as the similarity criterion. This is done only once at the lowest resolution of the wavelet transform. Only the matched blocks are carried for comparison to the next level. This drastically reduces the time needed for the detection process. This approach works even if the pasted region has undergone transformations like translation and rotation.