基于SURF的ROI区域的可逆水印算法  

Reversible Watermarking Algorithm Based on SURF for ROI Region

作　　者：夏兰婷 陈青[1] XIA Lan-ting;CHEN Qing(Signal and information processing,University of Shanghai for Science and Technology,Shanghai,200093)

机构地区：[1]信号与信息处理上海理工大学

出　　处：《软件》2020年第2期125-130,共6页Software

摘　　要：针对ROI水印算法对几何攻击的抵抗能力较弱的问题,本文提出一种基于SURF(Speeded Up Robust Features)特征点选取ROI,并分别嵌入中、低频子带的可逆水印算法,能有效抵抗几何攻击。该算法首先提取载体的SURF特征点,然后对载体图像进行5/3整数小波变换,筛选出感性兴趣区域(ROI)的低频系数和非感兴趣区域(ROB)的中频系数;水印经过置乱变换和抽样金字塔分解,水印分解后的近似子带嵌入感兴趣区域的低频子带中,残差子带嵌入非感兴趣区域的中频系数中。实验数据表明该算法能够抵抗常规的几何攻击和信号攻击,提取的水印相似度高,NC值均保持在0.89以上,具有良好的可逆性和鲁棒性。Aiming at the problem that ROI watermarking algorithm has weak resistance to geometric attacks,this paper proposes a reversible watermarking algorithm based on SURF(Speeded Up Robust Features)feature point selection ROI combined with medium and low frequency sub-bands,which can effectively resist geometry.attack.The algorithm first extracts the SURF feature points of the carrier,and then performs 5/3 integer wavelet transform on the carrier image to screen the low-frequency coefficients of the perceptual region of interest(ROI)and the IF coefficients of the non-region of interest(ROB).The watermark undergoes scrambling transformation.And the sampling pyramid decomposition,the approximate subband of the watermark decomposition is embedded in the low frequency subband of the region of interest,and the residual subband is embedded in the intermediate frequency coefficient of the non-region of interest.The experimental data show that the proposed algorithm can resist conventional geometric attacks and signal attacks.The extracted watermarks have high similarity and NC values are above 0.8,which has good reversibility and robustness.

关 键 词：SURF特征检测 ROI区域 抽样金字塔分解 可逆水印 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]