Video Forgery Detection Using Invariance of Color Correlation

Video forgery detection aims to trace a query video snip in a video file or database. It helps in detecting video copies for copyright protection and reduces storage redundancies. In this paper, we put forward an effective method based on the invariance of color correlation. The proposed technique first split each key-framework into nonoverlapping block For each block, the red, green, and blue color components are sorted according to their typical intensities, and make use of the percentage of the color correlation to build a frame feature with a minimum size. Finally, the ensuing video feature is made up of the successive frame features, which is verified to be robust beside most typical video content-preserving operations, together with geometric deformation, blurring, noise infectivity, contrast augmentation, and strong reencoding. The investigational results show that the projected method outperforms the accessible methods in the literature, as sound as the system based on the traditional color histogram.