节点文献
基于位平面统计相关的支持向量机图像隐写分析算法
Image Steganalysis Based on Bit-Plane Statistical Correlation Using Support Vector Machine
【摘要】 随着隐写技术的发展,隐写分析引起了越来越多研究人员的关注。通过对隐写前后图像位平面的相关性分析,提出一种空域LSB通用隐写分析算法。该算法只使用了32个特征统计量,具有较低的计算复杂度。实验结果表明,此算法对嵌入消息长度小于0 3bpp的载密图像的检测准确率超过90%,多次实验也表明此算法具有很好的稳定性。
【Abstract】 Steganalysis has attracted more and more researchers’ interest with the development of steganography technique. A universal LSB image steganalysis algorithm was presented by analyzing the statistical correlations between bit_planes and pixels in the same bit_plane. The algorithm only uses 32 statistics,and have a good computational complexity. The experimental results show that the detection rate is higher than 90% when the embedded message is less than 0.3 bpp. Many tests also show the algorithm has good stability.
【关键词】 隐写分析;
复杂度;
统计相关;
支持向量机;
【Key words】 steganalysis; complexity; statistical correlation; support vector machine;
【Key words】 steganalysis; complexity; statistical correlation; support vector machine;
【基金】 国家杰出青年基金资助项目(60325208);国家自然科学基金资助项目(60133020;60172067);广东省自然科学基金团队资助项目(04205407)
- 【文献出处】 中山大学学报(自然科学版) ,Acta Scientiarum Naturalium Universitatis Sunyatseni , 编辑部邮箱 ,2004年S2期
- 【分类号】TP309
- 【被引频次】8
- 【下载频次】196