Video Forensic of Fragmented Video Based on H.264/AVC Video Compression Standard

Abstract

Abstract-Traditional video forensics were just for complete video files, which aim at reconstructing the processing history of the video data and validating their origins or authenticity. They have obtained great achievements. However, we cannot always get complete video file in practice, sometime we only get part of it. In this paper, we came up with a method that can restore images of IDR frames from fragmented video files. After analyzing the format of MP4 and H.264/AVC Compression Standard, we proposed an algorithm to search for valid slices from fragment video files. With these slices, images of IDR frames can be restored by reconstructing sequence parameter set and picture parameter set. Test results show that images can be successfully restored in most cases except that FMO or data segmentation is adopted in the process of encoding.

Automatic Generation of H.264 Parameter Sets to Recover Video File Fragments

“…For example, in all videos the bit depths of luma and chroma samples are seen to be set to the same value, and the H.264 decoder implementation of FFMPEG video processing tool did not allow setting them otherwise. Finally, we must also note that in [13] only three parameters are identified to be crucial for decoding. Our observations on the design set in contrast show that the assumption of this work does not hold in practice.…”

“…Therefore, a more generalizable approach is to identify individual parameters needed to decode a file without a failure, as it can dramatically reduce the header search space. In fact, examining the parameters that comprise H.264 video headers, Sheng et al [13] identified 13 parameters and noted that only three of them are critical for decoding. Our approach, in this regard, is inspired by earlier work focusing on recovery of JPEG file fragments that determine decoding parameters to render the partial image from a block of JPEG coded data.…”