Video Object Removal Detection Based On Temporal Characteristics Clustering

In recent years, with the development and popularization of computer techniques,video editing software with complete function and easy operation has received moreand more attention. Therefore, simple digital video modifying has extended fromprofessional specialty of personnel to personal behavior of amateurs. Meanwhile,people start to doubt the authenticity of those videos published by common fans sincethis type of information resource is more personalized in producing and processing. Ifthese tampered videos are used in improper ways, unexpected negative effect will becaused. However, now there are still no effective detection systems and perfectsupervision measures. Thus, how to verify the authenticity and integrity of digitalvideos has become a key issue to maintain a sustainable development of digital videoindustry.Among a variety of multimedia video editing software, Imageineer MokeySystem quickly earn a lot of users because of its significant effect. Besides, it is verysimple to use. As to object removal editing, it only needs several steps to complete thewhole process, thus people can master it without a rich experience in video codec andvideo editing. In this paper a new approach is proposed based on the characteristics oftemporal sequences for detecting the copy-paste area of an object-removal-videoprocessed by Mokey.In our method, firstly a characteristics parameter is set when investigating theblock-level frames differences to extract the temporal characteristics as features forclassification. Then spatial contextual information is introduced into fuzzy clusteringto locate forged regions. According to the result the characteristics parameter iscalculated to repeat the previous two steps in order to get better location result.Eventually the morphological filtering is used to delete the isolated false locationblocks. Simulation results show that our method can achieves promising accuracy inidentification of video object removal and location of tampered area. Besides, ouralgorithm can adapt to the impact of MPEG4compression coding system.