A deep learning framework for copy-move forgery detection in digital images

Digital images have become widespread in modern life, and they can be modified and produced using an inclusive range of software and hardware tools. Since, digital image forgery can be extremely damaging, thus, understanding the detection and classification of authentic and forged images is of great significance. Without diminishing the importance of other types of forgeries, copy-move forgery (CMF) can be considered among the most widely utilized forgeries because of its ease of implementation. Nowadays, deep learning-based approaches are considered up-to-date for the classification and detection of image forgery owing to their improved accuracy and automated feature extraction skills. A deep learning CMF detection framework is proposed in this paper, which classifies images as authentic or forged using a contrast-limited adaptive histogram equalization (CLAHE) and convolutional neural network (CNN). The CLAHE algorithm makes the hidden features of the image visible, as some of them are hard to detect in CMF. The effectiveness of proposed structure is appraised using benchmark datasets: GRIP, MICC-F2000, IMD and MICC-F220. In terms of various performance metrics, the experimental study demonstrates the effectiveness of the proposed approach among other approaches. Also, the robustness of the proposed technique is demonstrated against several geometrical attacks like scaling, noise addition, JPEG compression, and rotation.