Improved SIFT-Based Copy-Move Detection Using BFSN Clustering and CFA Features

With the continuous development of multimedia technology, images in the real word and the Internet become more widespread, which brings unprecedented convenience to people's communication. However, the behavior of image tampering will brings disastrous consequences. Images could be forged using different techniques, and the most common forgery is copy-move forgery, in which a part of an image is duplicated and placed elsewhere in the same image. In this paper, we propose an improved scale invariant feature transform (SIFT)-based copy-move detection method, which combines broad first search neighbors (BFSN) clustering and color filter array (CFA) features. BFSN clustering algorithm is applied to detect multiple copied areas in tampered images. In order to distinguish original regions from tampered regions, we take CFA features into consideration to discover the inconsistency at the edges of the copied regions. Experiments demonstrate the efficiency of this method on different forgeries and quantify its robustness and sensitivity.