Tampering Detection and Restoration of Compressed Video
Bardia Azizian, Shahrokh Ghaemmaghami · 2018
This paper presents a method to detect tampering of video data and then restore an approximate version of its original contents in compressed H.264/AVC domain using watermarking. In the proposed scheme, a low resolution image from a number of video frames in certain time slots are embedded into the DCT coefficients of the other parts of the video which are adequately far from the reference frames. For detecting temporal/spatial tampering, the index of each frame/macroblock is embedded into itself as an authentication code. If a malicious tampering is detected at the authentication phase, the information used for restoration is extracted to recover the original contents in the altered areas. The extracted images are post-processed to enhance the image quality that could have been affected by channel errors or attacks. Our method provides appropriate transparency and robustness despite large volume of the payload to be handled. The results show that the BER of the watermark signal for recompression attack is about 4.5% on average under QP=24. The main achievement of the proposed watermarking system is the restoration of tampered areas, in addition to high accuracy in detecting malicious tampering in H.264/AVC domain.