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。
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.
出处
《中山大学学报（自然科学版）》
CAS
CSCD
北大核心
2004年第A02期17-20,共4页
Acta Scientiarum Naturalium Universitatis Sunyatseni
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)
关键词
隐写分析
复杂度
统计相关
支持向量机
steganalysis
complexity
statistical correlation
support vector machine
