关键词: 数字图像取证, 图像区域复制粘贴篡改, 不变矩

Abstract: Targeting a common means of tempering—the image region copy-move forgery, this paper proposes a detection algorithm based on moment invariant features. The detected image is divided into multiple same-sized overlapping blocks, whose moment invariant features and histogram features are extracted to compose a feature vector. The feature vector is sorted by dictionary to locate the copy-move region according to the predetermined similarity criteria. Experimental results show that the proposed algorithm performs better than the classical detection algorithm based on PCA（Principal Component Analysis） in terms of robustness against rotate operation and it can detect the 90° and 180° rotation accurately.

Key words: digital image forensics, image region copy-move forgery, moment invariant