| 基于预测模式相关性的HEVC隐写检测算法 |
| |
| 引用本文： | 盛琪,王让定,王斌,李倩,徐达文.基于预测模式相关性的HEVC隐写检测算法[J].光电子．激光,2017,28(7):766-772. |
| |
| 作者姓名： | 盛琪 王让定 王斌 李倩 徐达文 |
| |
| 作者单位： | 宁波大学 信息科学与工程学院,浙江 宁波 315211,宁波大学 信息科学与工程学院,浙江 宁波 315211,宁波大学 信息科学与工程学院,浙江 宁波 315211,宁波大学 信息科学与工程学院,浙江 宁波 315211,宁波大学 信息科学与工程学院,浙江 宁波 315211 |
| |
| 基金项目： | 国家自然科学基金(61301247,61300055)、浙江省自然科学基金(Z15F020010)、浙江省重中之重学科开放基金(xkx11405)和宁波大学优秀学位论文培 育基金(py2013005)资助项目 (宁波大学 信息科学与工程学院,浙江 宁波 315211) |
| |
| 摘 要： | 提出了一种基于预测模式相关性的高效视频编码(HEVC)隐写检测方法。首先定义帧内预测模 式相关性的概念；然后通过实验,分析 隐写前后帧内4×4亮度块中35种帧内预测模式 的相关性统计数据的变化情况；最后选择35个隐写前后变 化较大的相关性数据构成视频特征,并利用支持向量机(SVM)分类器进行隐写检测。实验表 明,本文检测 方法的特征维数低,计算复杂度小,对修改帧内预测模式的隐写算法具有良好的检测效果。
|
| 关 键 词： | 高效视频编码(HEVC) 预测模式 相关性 隐写分析 |
| 收稿时间： | 2016/7/17 0:00:00 |
Steganography detection algorithm based on the prediction mode correlation for HEVC |
| |
| SHENG Qi,WANG Rang-ding,WANG Bin,LI Qian and XU Da-wen.Steganography detection algorithm based on the prediction mode correlation for HEVC[J].Journal of Optoelectronics·laser,2017,28(7):766-772. |
| |
| Authors: | SHENG Qi WANG Rang-ding WANG Bin LI Qian and XU Da-wen |
| |
| Affiliation: | Ningbo University,College of Information Science and Engineering,Ningbo 315211, China,Ningbo University,College of Information Science and Engineering,Ningbo 315211, China,Ningbo University,College of Information Science and Engineering,Ningbo 315211, China,Ningbo University,College of Information Science and Engineering,Ningbo 315211, China and Ningbo University,College of Information Science and Engineering,Ningbo 315211, China |
| |
| Abstract: | At present,information hiding algorithm based on modified intra-pred iction mode has been proposed in the H.265/HEVC.The steganographic algorithm embeds secret information by modifying the intra-prediction mode.But there is no steganalysis method for H.265/HEVC from the available literature.This paper proposes a steganalysis method based on the correlation of intra-predicti on mode.The concept of the correlation of intra-prediction mode is defined firstly,then we an alyze the correlation of 35kinds of intra-prediction modes in the 4×4luma block before and after information hidi ng through abundant experimenl data.35correlation data which change a lot after information hiding are chosen to constitute the video features.Lastly,the video features will be sent to support vector machine (SVM) classifier to finish the detection of information hiding.Experimental results demonstrate that the algorithm has good detection ef fect while the feature dimension is low,and the computational complexity is small. |
| |
| Keywords: | |
|
| 点击此处可从《光电子．激光》浏览原始摘要信息 |
|
点击此处可从《光电子．激光》下载全文 |
|
