ARCHIVES

Copy-move image forgery is a widely practised digital manipulation technique in which one or more regions of a digital image are duplicated and repositioned within the same image to conceal objects, fabricate evidence, or create a misleading visual narrative. This paper presents a comprehensive hybrid forensic detection system that integrates four complementary algorithmic stages: (i) Discrete Stationary Wavelet Transform (DyWT) for noise-robust frequency-domain preprocessing; (ii) Scale-Invariant Feature Transform (SIFT) for the extraction of rotation- and scale-invariant keypoint descriptors; (iii) a Generalised 2-Nearest Neighbour (G2NN) self-matching algorithm for identifying statistically self-similar descriptor pairs within the same image; and (iv) agglomerative hierarchical clustering with cluster-pair verification to suppress false positives. Experimental evaluation on five standard benchmark datasets (MICC-F220, CoMoFoD, and COVERAGE, COLUMBIA, and CASIA v2) demonstrates a mean detection accuracy of 94.2% with an F1-score of 94.0%, surpassing comparable block-based and keypoint-based state-of-the-art approaches. The system is delivered as a cross-platform PyQt6 desktop application supporting background-threaded execution, live progress monitoring, and automated forensic report generation in PDF and TXT formats, making it immediately deployable in digital forensic investigation contexts.