An Improved Technique for Identification of Forgery Image Detection Using Clustering Method

In forensics, the detection of forgery images is very important. In this method of forgery, the same image of the region can be able to copied and pasted. The keypoint-based method was advertised to reveal ineffective forgery detection against different attacks like geometric transformations because the number of keypoints is less in small regions. To overcome these problems, the following methods are involved. In the first stage, we extract many keypoints by reducing the low enhancement by the CLAHE algorithm. After extraction, there will be some keypoint matching problems to solve this problem the FANN algorithm is used. At the final stage, the tampered region is localized, and in order to overcome FPR, an iterative localization technique called the Density-based clustering algorithm and GORE method is used. By using the software MATLAB R2017a, the performance of experimental results is evaluated and accuracy of 93% by image and pixel level.