Image Resampling Forensics: A Review on Techniques for Image Authentication

Abstract

The raw image can be corrupted in part or as a whole; thus it is essential to identify the type of image manipulation technique employed and region of tampering on the infected image. Initially, individual handcrafted modified images were used to recognize the tampering contained in the image, but in the actual world, one image can be edited using a variety of image manipulation schemes. However, at present, numerous tampering schemes are probed on an image, followed by post-processing to remove the traces and signs left by the tampering operation. This process makes the detection of tampering difficult and hassle task even for efficient detector. Image resampling is a common manipulation of particular forensic importance since resampling is performed at any stage of processing such as resize, rotation, or affine transformation. The forensics of resampling plays a prominent role to detect tampering in images. So resampling detection is a significant tool that provides the forensic clues about the forged image. Generally, resampling is used to reconstruct the history of an image that will be helpful in detailed forensic analysis. In this chapter, a detailed review on various recent image manipulation detection techniques of image authentication using resampling forensics is studied.