Passive Forensics Of Video Frame-Rate Up-Conversion

The prevalence of video acquisition devices and powerful processing software have given rise to large amounts of doctored videos with no obvious traces.The traditional concept of 'seeing is believing',which is subversion.Frame-based operation is one of the specific operations of video forgery including frame duplication,frame deletion,frame-adding and frame reorganization.What's more,video frame-rate up-conversion(FRUC)is one of the specific temporal-domain operations.It used to improve the motion continuity of video,but can also be used for the purpose of malicious tampering.Therefore,a great demand for automatic forgery detection algorithms that can identify the trustworthiness of a candidate video.Until now,The research of video FRUC forensics has just begun,Therefore,passive forensics against video frame rate up conversion both technical challenges and great application potential.This paper focused on forgery detection against FRUC,through the statistical characteristics of the video will be varied after FRUC.Research on the corresponding passive Forensics,and the innovations and contributions are as follows:Firstly,a passive forensics approach based on periodic properties of edge-intensity is proposed to detect the possible video FRUC operation.After computing the edge intensities of every frame,Kaufman adaptive moving average(KAMA)is exploited to define an adaptive threshold to distinguish the interpolating frames by FRUC from the original frames.Moreover,the original frame-rate of up-converted video can be inferred.Experimental results show that the proposed approach is not only effective for simple frame repetition and linear interpolation,but also valid for advanced FRUC techniques such as motion compensated frame interpolation(MCFI).The detection accuracy is up to 94.5%on average.Its computation is simple as well.Secondly,a passive forensics approach based on the image-sharpness and correlation coefficients of frame-difference is proposed to detect the possible video FRUC operation.Firstly,the method uses the morphological gradient to calculate the sharpness of the frames,then determining the video whether the FRUC or not according to the abnormal sharpness curve.Moreover,for the judgment process,it may be emergence some test video can't be accurately determined,but the suspicion videos possible are tampered with by FRUC,then the need to introduce another type of forensics features just like frame-difference of the correlation coefficient.The suspicion videos are detected again by the frame-difference of the correlation coefficient and finally locate position of the tampered frames.Experimental results show that the average detection accuracy of proposed algorithm is high,and the detection of the operation of frame-rate up conversion tamper works well.