CAJ | 학술논문

针对视频水印在帧内篡改检测方面定位精度的不足，通过压缩感知对MPEG-4（Moving Picture Experts Group-4）视频内容的特征表示，提出一种新的视频水印生成方法及其帧内篡改检测算法。该算法由压缩感知DCT （Dis-crete Cosine Transform ）测量矩阵对I-VOP（Intra-Video Object Plane）图像提取 U、V特征参数，生成基于内容的压缩感知视频水印数据并嵌入到图像 Y分量的DCT中高频系数中实现帧内篡改检测。实验结果表明，与Hash视频水印算法比较，压缩感知视频水印数据具有更好的恢复能力，且水印算法对视频帧内篡改定位精度更高。
침대시빈수인재정내찬개검측방면정위정도적불족，통과압축감지대MPEG-4（Moving Picture Experts Group-4）시빈내용적특정표시，제출일충신적시빈수인생성방법급기정내찬개검측산법。해산법유압축감지DCT （Dis-crete Cosine Transform ）측량구진대I-VOP（Intra-Video Object Plane）도상제취 U、V특정삼수，생성기우내용적압축감지시빈수인수거병감입도도상 Y분량적DCT중고빈계수중실현정내찬개검측。실험결과표명，여Hash시빈수인산법비교，압축감지시빈수인수거구유경호적회복능력，차수인산법대시빈정내찬개정위정도경고。
In order to overcome the drawback of localization accuracy for intraframe tampering detection based on video wa-termarking ,this paper proposes a novel video watermarking generation and intraframe tampering detection algorithm based on feature representation of MPEG-4 (Moving Picture Experts Group-4 ) video content by compressed sensing (CS ) .First ,U and V ,feature pa-rameters of I-VOP(Intra-Video Object Plane) frame ,are extracted utilizing projection of DCT (Discrete Cosine Transform) mea-surement matrix of the CS ,generate the CS-watermarking of the video content ,and embed the CS-watermarking data into the coeffi-cients of low frequency and high frequency of DCT domain in the Y component of image .Compared to Hash-watermarking generat-ing algorithm ,experiment results show that the CS-watermarking algorithm has better data recovery capability and localization accu-racy for video intraframe tampering detection .