Enhanced Techniques for Detecting Copy-Move Forgery Using Machine Learning Algorithms

Abstract

One of the most important tasks in digital forensics to find instances of modified content is the detection of copy-move forgery (CMF) in videos. Copy-move forgery includes taking a section of a video, pasting it into another movie, and then hiding or changing that section. As a consequence of advancements in network technology, low-cost multimedia devices, intelligent image or video editing software, and broad adoption of digital multimedia coding standards, the number of applications for digital multimedia has significantly risen in recent years. Establishing if a video is legitimate or not is one of the trickiest areas of video forensics. This may be a crucial responsibility when recordings are used as primary evidence to influence decisions, such as in a court of law. Therefore, we provide a novel machine learning-based copy-move forgery detection technique in this research. Weiner filter is first used to gather and pre- process video data. The pre-processed video data are then segmented using a threshold-based technique to image segmentation. Finally, we suggest a novel integrated stochastic random neighbouring approach (ISRNA) for categorizing videos. Our suggested technique is compared and contrasted with traditional ways to demonstrate the efficacy of the suggested method. Results from experiments show that our suggested strategy performs better than traditional ways.