Efficient detection of copy-move forgery using polar complex exponential transform and gradient direction pattern

• Published in: Multimedia tools and applications Vol. 82; no. 7; pp. 10061 - 10075
• Main Authors: Babu, S. B. G. Tilak, Rao, Ch Srinivasa
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.03.2023; Springer Nature B.V
• Subjects: 1225: Sentient Multimedia Systems and Universal Visual Languages; Computer Communication Networks; Computer Science; Data Structures and Information Theory; Digital imaging; Feature extraction; Forensic sciences; Forgery; Histograms; Multimedia Information Systems; Performance measurement; Recall; Special Purpose and Application-Based Systems; Copy move forgery detection; Gradient direction pattern; Image forensics; Computer crimes; Polar complex exponential transform
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-022-12311-6

Evidence plays a vital role in image forensics. If evidence is an image, then its authenticity verification is the key to image forensics. One of the common forgeries in digital images is Copy-Move Forgery, which happens in a single image in which some portation of the image is copied and pasted in the same image. Copy Move Forgery Detection has demand in legal evidence, forensic examination and many more areas. The proposed method starts with the conversion of a grey image into overlapping blocks. Rotationally invariant stable Polar Complex Exponential Transform features are obtained from each overlapping block. The extracted feature dimensionality is further reduced using the Gradient Direction Pattern histogram. The similarity is identified among these histogram feature matrix rows. False matches are eliminated with the help of the windowing technique and morphological operators. The performance of the proposed method is calculated in terms of recall rate, precision, and F1score. The testing results are outstanding, even when the suspected image has been subjected to post-processing assaults; the recall rate is the highest in the literature, and the remaining performance metrics are likewise excellent.