Detection of Copy Move Forgery Using Rootsift and Feature Point Matching

Abstract

Abstract The popularity of the internet has grown recently, making it simpler to broadcast and purchase digital content. This popularity suggests that counterfeiting has increased the susceptibility of multimedia to modification. A copy-move forgery is one of the most popular techniques for digital photo modification. Key point-based detection systems have successfully exposed copy-move forgery as a result of various digital assaults. On the other hand, these methods are useless when forgeries only involve tiny, smooth areas with a small number of crucial spots. A multi-scale feature matching based on the copy-move forgery detection technique was put into place to solve these problems. I believe the copy move method has some potential as well. Reduce the intensity threshold and rescale the supplied image to produce a reasonable number of key points in order to address those concerns in the proposed system initially. Then, a special multi-scale matching method is developed using scale clustering, overlapping grey scale clustering, and group matching modules. Present a novel iterative localization strategy that locates the forged region by utilising the resilience properties of each key points as well as its colour information. According to the experimental results, the suggested work's copy move forgery average is 1.375, compared to the existing system's average of 2.75, with standard deviations of 0.72 and 1.44, respectively. As a result, we can say that the proposed approach is more accurate and efficient than existing methods by more than 99.9%.