Research On Surveillance Video Recompression Detection Algorithm For VVC Standard

Video surveillance systems play an important role in maintaining public safety and social stability,and with the growing demand for high quality video services,HD surveillance cameras have become widely used in our daily lives.With this comes a large number of HD surveillance videos.However,the increasing power of video editing software and the development of deep tampering technology have made it easier and easier to tamper with surveillance video.Therefore,how to ensure the authenticity of surveillance video content has become an important issue of concern for researchers.In order to solve the problem,various digital video forensic techniques are widely studied,among which,video recompression forensic technology is a key technology in the field of digital video forensics for video authenticity identification.The tampered video must have been compressed twice or more,so the authenticity of the video can be judged by detecting the number of compressions.Video recompression forensics technology for a variety of video coding standards,the latest generation of international video coding standards for versatile video coding(VVC)standard.The current video recompression forensics technology is oriented to the encoding standard before VVC.And the current video recompression forensics technology only considers the coding information of luminance domain,ignoring the coding information of chroma domain.Based on this,this paper is oriented to VVC standard,for surveillance video content,from the consideration of the coding information of video chroma domain and luminance domain,while considering the massive amount of data of high-definition surveillance video,dedicated to improve the accuracy of surveillance video recompression detection,while reducing the complexity of the algorithm.The research done in this paper is described as follows.(1)Based on the hybrid coding principle of VVC standard,combined with the Lagrangian rate distortion optimization technique,the basic bitstream characteristics in VVC video bitstream,which are closely related to the compression time,are analyzed and determined from the dimensions of chroma domain and luminance domain in surveillance videos.The basic bitstream characteristics include the division types and prediction patterns of the chroma domain and luminance domain coding units(CU).Finally,it is experimentally verified that the above features are closely related to the compression number,and the more advanced the compression number is,the closer the connection is.Therefore,the partitioning type and prediction mode of CU in the chroma domain and luminance domain can be used as the basic bitstream characteristics to detect the number of video compressions.(2)To address the problem that the current video recompression forensics method does not consider the coding information of chroma domain and the forensic accuracy is low,we propose a surveillance video recompression forensics method for the information fusion of chroma domain and luminance domain in VVC standard,abbreviated as CLF-SVRF,and consider the demand of video surveillance applications for recompression forensics method with low complexity,and build a low-complexity method based on the partitioning type and prediction mode of CU mode to effectively fuse the information of chroma domain and luminance domain.The advanced the basic bitstream characteristics with low complexity are constructed based on the partitioning type and prediction mode of CU to effectively integrate the information of chroma domain and luminance domain.Finally,the advanced the basic bitstream characteristics are fed into a linear classifier in the field of machine learning to complete the recompression forensics of surveillance video.The experimental results show that the proposed CLF-SVRF method improves the accuracy of surveillance video recompression forensics by 13.10% on average and reduces the recompression forensics time by 47.42% on average compared with the current advanced methods.In summary,the surveillance video recompression forensics proposed in this paper considers the new technology of VVC standard,which starts from the chroma domain and luminance domain dimensions of surveillance video,and determines the authenticity of surveillance video by detecting the compression times of the video.The experimental results show that the algorithm in this paper significantly improves the accuracy of recompression forensics of surveillance video,and at the same time considers the massive data of HD surveillance video and reduces the time complexity in constructing advanced bitstream characteristics.