ABSTRACT
This chapter focuses on copy-move and splicing as they are the most common forms of image tampering techniques. It discusses the various classifications of them based on the detection mechanism like keypoint based, and texture based. Copy-move forgery/tampering detection essentially involves looking for identical regions in the image based on some kind of matching. The Speeded Up Robust Features consists of a keypoint detector and descriptor. The keypoints are generated using Fast-Hessian Detector which is based on the Hessian matrix for a particular image pixel. The keypoint descriptor is formed from the wavelet responses in a particular neighborhood of the keypoint and Haar wavelet responses are used for orientation assignment. Localization and detection of copyrotatemove manipulations using Zernike moments is discussed over small overlapping image blocks. A multi-radius polar complex exponential transform is used by extracting rotational invariant and multi-scale features.
