A Two-Stage Cascaded Detection Scheme for Double HEVC Compression Based on Temporal Inconsistency

Abstract

Nowadays, verifying the integrity of digital videos is significant especially for applications about multimedia communication. In video forensics, detection of double compression can be treated as the first step to analyze whether a suspicious video undergoes any tampering operations. In the last decade, numerous detection methods have been proposed to address this issue, but most existing methods design a universal detector which is hard to handle various recompression settings efficiently. In this work, we found that the statistics of different Coding Unit (CU) types have dissimilar properties when original videos are recompressed by the increased and decreased bit rates. It motivates us to propose a two-stage cascaded detection scheme for double HEVC compression based on temporal inconsistency to overcome limitations of existing methods. For a given video, CU information maps are extracted from each short-time video clip using our proposed value mapping strategy. In the first detection stage, a compact feature is extracted based on the distribution of different CU types and Kullback–Leibler divergence between temporally adjacent frames. This detection feature is fed into the Support Vector Machine classifier to identify abnormal frames with the increased bit rate. In the second stage, a shallow convolutional neural network equipped with dense connections is designed carefully to learn robust spatiotemporal representations, which can identify abnormal frames with the decreased bit rate whose forensic traces are less detectable. In experiments, the proposed method can achieve more promising detection accuracy compared with several state-of-the-art methods under various coding parameter settings, especially when the original video is recompressed with a low quality (e.g., more than 8%).

A Review on HEVC Video Forensic Investigation under Compressed Domain

“…13 and some P frames to be reencoded as P frames (P-P frames). When compared to neighbouring P-P frames, relocated I-P frames possess abnormal characteristics in basic coding elements, such as variation in statistics of CU, PU and TU; disproportion of intra, inter and skip units; and disparity in Intra modes has been observed [26][27][28][29][30][31]. Authors in [26] counted the number of intra, inter and skip PUs and designed a feature vector named, SN-PUPM, by taking absolute difference with adjacent three frames.…”

“…These models are capable of automatically learning features from visual data. The authors in [30,31] presented approaches by using a convolutional neural network. Authors in [30] proposed an attention-based two-stream ResNet (ATResNet) network integrated with LSTM model.…”

“…He et al [31] examined that low order statistics applied on the distribution of CU types, including intra, inter and skip CUs, can efficiently predict relocated I-frames under increased bitrate. Whereas these features are insufficient to detect relocated I-frames encoded with decreased bitrate.…”