| In the digital multimedia age, video forgery is getting easier and easier with the popularity of digital cameras and video editing softwares. Digital synthesis is a common means of nonlinear video production, and one of the important functions is keying technology. The technology is simple with realistic effects, and has become popular in film stunts field. However, criminals will also take use of keying technology in video forgery and tampering. Once the false video is used in fields such as news, scientific, judicial, military and political, there’ll be serious social consequences. Therefore, how to identify the digital image/video integrity and authenticity has become a hot research field of information security.Each digital camera contains relatively stable residual noise due to the imperfection of imaging sensors, which is a unique sensor noise pattern and serves as a unique identification fingerprint. This paper focuses on detecting digital video forgery based on pattern noise. At first, the basic theory and method of digital video forgery detection with pattern noise is discussed. Then according to the references, the pattern noise of specific cameras is extracted using the wavelet domain image denoising algorithm by wiener filter. After that, chroma keying video forgery can be exposed in the Fourier transform domain by the difference of noise patterns between original videos and the forged videos. The results of the experiments show that, with different keying materials and at different bit-rates, this detecting algorithm can still work efficiently and accurately. |
PDF Full Text Request