模拟退火的安全数据隐藏算法  

Secure data hiding algorithm based on simulated annealing

作　　者：刘光杰[1] 戴跃伟[1] 王金伟[1] 王执铨[1] 

机构地区：[1]南京理工大学自动化学院,江苏南京210094

出　　处：《光学精密工程》2007年第5期791-797,共7页Optics and Precision Engineering

基　　金：国家自然科学基金资助项目(No.60374066);江苏省自然科学基金资助项目(No.BK2004132)

摘　　要：以数据嵌入前后JPEG图像DCT系数统计分布的平均相对熵(ARE)作为信息隐藏的安全性度量,并以此作为优化指标,给出一种带调整的量化嵌入算法,通过模拟退火选择对统计分布破坏最小的调整变量。实验结果表明:对于五幅标准测试图像,本文算法较其他四种算法在容量上平均提高了1.86倍;而在平均相对熵指标上平均降低了38%。该算法不但具有较高的数据容量,并可有效地补偿数据嵌入导致的统计分布的变化,实现算法在统计意义上的安全性。Taking the Average Relative Entropy（ARE） of DCT coefficients distribution before and after data hiding as the measures of security and the security optimization obiect, a optimal adiustment variables in the quantization embedding algorithm are chosen in simulated annealing. The experimental results based on the five standard test images show that comparing with other four algorithms the capacity is improved by 1.86 times in average, and the ARE object is decreased by 38% in average.And they also show that the proposed method can decrease the statistical distribution changes of DCT coefficients to keep the security in statistical sense with larger steganographic capacity.

关 键 词：数据隐藏 隐写 统计安全性 模拟退火 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]