AUTOMATIC FORGED SCENE DETECTION IN ADVANCED VIDEO USING COMBINED MPEG-2 AND OPTICAL FLOW FEATURES

Multimedia forensics has been attracted more attention in recent times due to extensive growth of floating multimedia contents in social media. Passive multimedia forensics plays a key role in identifying forgery contents from the multimedia contents. Furthermore, identifying such forged content in the video is always a hard task. Research on video forensics is still emerging and requires an exhaustive research for advancement. In this research we recommend a novel feature based technique to find forged scene from the video. The process involved in this work is fully automated and uses the combination of MPEG-2 and Optical flow features. Optical flow features is robust in localizing the pixel flow for per frame analysis and motion vectors are spawned from the target video frames. An automatic forge scene detection algorithm is also formed using block matching algorithm. Block matching algorithms was designed to calculate displacements. Further we regard that the proposed combined feature extraction method are advancement to extract forensic features from the motion vectors in advanced video. The entire framework is built using MATLAB R2018a version running in Intel Xeon processor with 64 GB RAM. An experimental result also confirms that the proposed method is better in terms of results and capable of detecting forged frames in the advanced video.