Implementation of Image/Video Copy- Move Forgery Detection Using Brute-Force Matching

The images and videos can be easily forged using low cost editing tools and software's. This paper, introducing a new method called brute-force matching for detecting image and video copy-move forgery. In image forgery detection, divided the input image first into nonoverlapping and irregular blocks adaptively using adaptive oversegmentation method. Then feature points are selected from each blocks using scale invariant feature transform. To identify the copy-move forgery, brute-force matching is used for matching the selected features with one another. To detect accurate forgery regions, morphological operation is also used. In video copy-move forgery detection, object tracking method is used instead adaptive oversegmentation. First step in video forgery detection is dividing the input copy-move video into frames and convert each frame to gray scale images. The frame keypoints and descriptors are detected using scale invariant feature transform. Then the detected descriptors are matched with one another in the frame using brute-force matching. Finally, morphological operation is applied for getting an accurate result. The proposed method is implemented on Raspberry pi 3 model b using python-OpenCV platform.