基于小波域纹理分析的图像自适应隐藏算法  

An adaptive data hiding algorithm in wavelet domain based on texture analysis

作　　者：王媛媛[1] 俞一彪[1] 

机构地区：[1]苏州大学电子信息学院,江苏苏州215021

出　　处：《电路与系统学报》2010年第1期96-101,共6页Journal of Circuits and Systems

摘　　要：提出了一种以彩色图像为载体的变换域自适应嵌入方法,根据小波变换后的低频子带保留有原图像大部分能量和主要的视觉特性,利用熵值、标准差进行纹理分析并标记出平滑、边缘、纹理区域,再由各区域的视觉阈值决定不同的信息嵌入深度。预处理时将小波低频系数末比特清零,实现信息隐藏的盲恢复。此变换域方法对复杂纹理性的图像嵌入率可达13.89%,且常规的鲁棒性实验表明本文方法可获得较好的恢复图像。A new adaptive image hiding method is proposed. According to that the low-frequency sub-band of DWT retains a majority of energies and image visual characteristics, the low-frequency part is classified as smooth, edge and texture by entropy and standard deviation, and different regions are assigned with different hiding bits, which will increase the hiding capacity. Resetting low bits in DWT low-frequency as pretreatment makes the blind detection accomplished. The embedding rate of complex texture image is up to 13.89%, and experimental results show this method is feasible and routine to attacks.

关 键 词：彩色图像 信息隐藏 整数提升小波 自适应 盲检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP309[自动化与计算机技术—计算机科学与技术]