Detecting frame duplication based on spatial and temporal analyses

A passive scheme is developed to achieve frame duplication detection for videos in the paper. The proposed scheme is a coarse-to-fine approach and composed of candidate clip selection, spatial correlation calculation, and frame duplication classification. To screen duplicated candidates in the temporal domain, the histogram difference of two adjacent frames in RGB color space is adopted. To evaluate the similarity of image content, a block-based algorithm is used to measure spatial correlation of each corresponding frame between the query clip and the candidate one. Based on the spatial and temporal analyses in the temporal and spatial domains, we can localize the duplicated clips. The experimental results show that our scheme can not only detect but also localize duplicated clips well. In addition, the proposed scheme outperforms an existing method in terms of false alarm, missed detection, and accuracy.