A new video forgery detection approach based on forgery line

Abstract

In recent years, we have continually encountered multimedia files in daily life, as evidence in courts, medical records in hospitals, etc. Many multimedia-editing software tools have been developed in parallel to the widespread usage of multimedia files. Thus, forgery operations can be committed by anyone with this software, even if s/he does not have any skill in this field. Authentication of the originality of multimedia files has recently become a popular topic. In this work, we propose a new video forgery detection approach to detect forged frames with better execution and detection capability. Features are extracted from the frames and their correlations are represented as a correlation image. This method investigates a line on the correlation image to determine the forgery operation. Then two new procedures (shrinking/expanding) are applied to the detected line to determine the exact location of the forgery. The experimental results indicate that the proposed method can detect forgery operations with better execution time and detection capability when compared to similar works in the literature.