NOVEL S4 FILTER APPROACH FOR COPY-MOVE VIDEO FORGERY DETECTION

Abstract

Video forgery detection is a challenging task nowadays due to fake forwarding through several social media apps. Especially copy-move type of attack is broadly in practice to tamper the original contents of the video or an image. In this paper, we have proposed a novel approach for the detection of copy-move attack in passive blind videos. We have inspected the forged video for detection of two scenarios, such as a single object is copied and pasted at a single time and a single object is copied and pasted multiple times within the same video. The algorithm proposed combines two novel techniques for object detection and copy-move detection respectively. Object detection and tracking are carried out based on a new contour-based approach and a copy-move attack is implemented by a novel S4 filter-based algorithm. These filters are- Slope (S1), Span (S2), and Size (S3) Similitude Index (S4). The proposed algorithm is tested and implemented for both standard and custom dataset for the detection of copied objects by tracing the exact frame number of the forged objects. With the combining effect of both techniques, the algorithm gives faster and accurate results than the state-of-art techniques.