Abstract
In the realm of digital multimedia analysis, combating picture fraud stands as a critical endeavor, given the prevalence of manipulation facilitated by contemporary multimedia creation tools. The ease of copying, pasting, and tampering with digital images poses a significant challenge, altering the reality of the original image and constituting an illegal operation. While existing pixel- and transform-based methods have exhibited superior detection and estimation capabilities, they are not without limitations. This research proposes a novel texture-based approach for the identification of fake images, aiming to overcome the drawbacks of current methodologies. The suggested method proves effective in terms of detection ratio, offering a promising solution to the complexities introduced by contemporary multimedia manipulation. The extraction of texture features is accomplished using a discrete wavelet transform algorithm, enhancing the robustness of the forgery detection process. A crucial aspect of the proposed approach is the block creation method, implemented through the partition clustering approach. This facilitates the creation of blocks for both genuine and manipulated images, contributing to a comprehensive analysis. To validate the efficacy of the suggested method, extensive testing is conducted utilizing the well-known MFIC2000 dataset and MATLAB software. The outcomes of this research not only shed light on the advancements in texture-based forgery detection but also provide a practical and efficient solution address the challenges posed by picture fraud in the contemporary multimedia landscape. The proposed methodology contributes significantly to the evolving field of multimedia forensics, enhancing the arsenal of tools available for detecting and mitigating digital image tampering.
Keywords: Cluster segmentation, DWT, texture, and image tampering.