论文部分内容阅读
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。
Powerful and easy-to-use video editing software can subject digital video to various forms of tampering and the authenticity and integrity of the video can not be guaranteed. Dual compression is a necessary condition for video tampering, and double compression testing is an important aid to video forensics. By analyzing the discrete cosine transform (DCT) coefficient variation caused by quantization error during compression, a double-compression video compression algorithm based on high-efficiency video coding (HEVC) with different quantization parameters is proposed. By using DCT coefficient histogram and adjacent DCT coefficients A 22-dimensional joint feature set is constructed for the statistical properties of parity pairs. Finally, the feature set is classified and identified by Support Vector Machine (SVM). Experimental results show the effectiveness of the proposed algorithm.