Recapture Video Detection Based On Feature Analysis

With the rapid development of science and technology,and with the widely used of low cost digital electronic products(such as a digital camera,smart phones with HD camera and HD video camera,etc.),digital video has become an important part of our life.Under the surface of the prosperous,however,the development of science and technology also brings unfavorable factors,the film piracy,which not only affects the immediate interests of the copyright owner,at the same time,poor pirated digital multimedia products will also bring bad audio-visual experience to the user.By far,the most common method is to use with encryption algorithm such as digital copyright protection to ensure the safety of the transmission path,but for pirates directly use camera in front of a screen display to recapture video,post method has no effect.In this paper,we propose two recapture video detection algorithms based on dynamic shoot and static shoot video respectively.From the perspective of the dynamic shoot video,we presents a recapture video algorithm based on the consistency of velocity field.The algorithm is mainly using the movement motion of original video is orderly.And all recapture hand-held video will bring jitter character to make video trend movement disorder.According to this principle,the algorithm can effectively judge whether video through recapture process.The algorithm's accuracy and efficiency is very high.From the perspective of the static shoot video,we presents a recapture video algorithm based on the profile dictionary matching degree.The basic principle of the algorithm is that the recapture video frame's distortion is heavier than the original video,the sharpness will be lower,and will muchblurrier than the origin one.According to this phenomenon,the algorithm build a dictionary for image edge,through calculating the matching degree of image edge and dictionary to determine whether a video through recapture process.