Image tampering detection using dynamic histogram equalization based LIOP features and novel scaled K-means +  + clustering

Digital images are one of the most regularly used means of storing personal memories and legal proofs. Content modification of digital images is becoming more and more common due to highly advanced image processing tools and techniques. However, digital image modification tools are used to change or hide the digital image content. Copy-move forgery (CMF) is an extensively used image alteration technique in which similar content in that particular digital image alters the contents of the digital image. Many detection methods for CMF have been proposed by various researchers. The keypoint-based CMF detection (CMFD) approach is considered robust and effective. However, the keypoint-based (KB) CMFD approaches are not able to produce satisfactory accuracy, if the tampered regions are small. A novel KB-CMFD approach is presented in this article to detect the small tampered regions in digital images. The CMFD approach proposed in this article is based on the dynamic histogram equalization (DHE), local intensity order pattern (LIOP) as a feature descriptor, and a novel scaled K-means + + (sK-means + +) clustering to detect small regions. The passive KB-CMFD proposed approach performance is evaluated on standard datasets (the CoMoFoD, the MICC-F220, and the defecto MSCOCO (synthetic)). The proposed passive KB- CMFD perform well compared to the other state-of-the-art CMFD approaches.