Double compression detection based on local motion vector field analysis in static-background videos

摘要: A novel double compression detection method in static-background videos is proposed.Local motion vector field analysis used to obtain segmentation of video contents.Modification based on local strengths extract robust fingerprints.The proposed has more performance than several state-of-the-art methods. Videos captured by stationary cameras are widely surveillance and conference. This kind often static or gradually changed background. By analyzing the properties videos, this work presents a approach detect MPEG-4 videos. For given suspicious video, segment background regions each frame. According backgrounds strength foregrounds, modified prediction residual sequence calculated, which retains fingerprints compression. After post-processing, GOP estimation results obtained applying temporal periodic final feature sequence. Experimental have demonstrated better robustness efficiency comparison Besides, various rate control modes.