基于DCT系数统计特性的HEVC视频双压缩检测算法

功能强大和使用简易的视频编辑软件可能会使数字视频遭受到各种不同形式的篡改,视频的真实性和完整性无法得到保证。双压缩是视频篡改的必要条件,双压缩检测则是视频取证的重要辅助手段。通过分析压缩过程中由量化误差引起的离散余弦变换(DCT)系数变化,提出了一种不同量化参数下的高效视频编码(HEVC)视频双压缩检测算法,利用DCT系数直方图和相邻DCT系数对奇偶组合统计特性构造22维联合特征集,最后将特征集用支持向量机(SVM)进行分类识别。实验结果证明了本文算法的有效性。

Detection of double compression in HEVC videos based on the statistical characteristic of DCT coefficient

With the wide spreading of powerful and easy-to-use video editing sof tware,digital videos are exposed to lots of different tampering kinds which destroy the authenticit y and integrity of videos.Since double compression is necessary to video tampering,the detection of double compression is an important supplementary means for digital video forensics.By analyzing changes in discrete cosine transf orm (DCT) coefficients caused by quantization,an algorithm of detecting double compression for high efficiency video coding (HEVC) videos i n different quantization parameters is proposed,which employs 22-dimensional union feature set consisting of the D CT histogram and the parity group statistics of DCT coefficient-pair, and then the feature set is sent to support vector machine (SVM) classification to distinguish double compression videos from single compression videos.Experimental verify show the effect iveness of the algorithm.