加密域视频可逆数据隐藏关键技术研究

为适应云计算数据安全和隐私保护需求，敏感视频以加密形式存储在云端。在很多应用场景，云端需在加密视频中嵌入一些可擦除附加消息，如视频标注或认证数据。为此，本项目开展加密域视频可逆数据隐藏关键技术研究，主要内容包括：1)研究加密域可逆数据隐藏"嵌入率--失真"函数，构建联合"载体无损压缩--数据嵌入"的近似最优可逆数据嵌入编码；2)针对直方图移位的可逆数据嵌入技术,研究有效的H.264/AVC加密方案，利用DCT残差系数和运动矢量差分构造加密载体的高集中统计量直方图；3)研究加密统计量直方图移位嵌入方案，将近似最优可逆数据嵌入编码引入到直方图移位嵌入中，以提高算法的嵌入效率；4)挖掘可逆数据隐藏引起的H.264/AVC视频帧内失真漂移机理，研究可逆失真漂移补偿技术。本项目的研究将为加密域可逆数据隐藏提供新思路和技术基础，预计研究成果可应用于云计算环境下的视频数据标注、版权保护及篡改检测等领域。

In order to meet the demand for data security and privacy protection in cloud computing, sensitive videos should be stored in the encrypted form in cloud. In some applications, cloud service providers hope to append some additional message, such as video notation or authentication data, in a reversible manner. In order to do this, research on key technologies of reversible data hiding for encrypted videos is carried out. The main contents include: 1) Explore and research the "embedding rate--distortion" function of reversible data hiding in the encrypted domain, and establish the near-optimal coding structure by using joint carrier lossless compression and data embedding. 2) Research effective H.264/AVC encryption schemes for the histogram-shifting based reversible data hiding technique, and establish the highly centralized statistical histogram of encrypted videos by using DCT residual coefficients and motion vector differences. 3) To improve embedding efficiency, the histogram-shifting based reversible data hiding technique in the encrypted domain with the combination of near-optimal coding structure should be investigated. 4) Reversible distortion drift compensation technology is further studied based on the mechanism of frame distortion drift caused by reversible data hiding in H.264/AVC video. The project will provide new ideas and theoretical basis for reversible data hiding in the encrypted domain. The research results can be applied to multimedia data annotation, copyright protection and content authentication in cloud computing.