摘要
针对基于分块思想的图像定位算法定位精度不高和恢复性能不好的问题,本文提出了一种基于精细认证和迭代补偿机制的图像认证与恢复算法。在认证算法中,包含基本认证和精细认证,融合基本认证和精细认证这两次认证的结果,并根据连通篡改区域中像素个数的多少进行优化,得到定位精度较高的定位结果。在恢复算法中,考虑到可嵌入的信息容量,提出利用迭代补偿机制恢复图像中被篡改的像素。理论分析和实验结果表明:本文提出的算法对于局部发生篡改的图像,能提高其定位的精度,而且恢复的性能也得到了提高。
To enhance the accuracy of tamper localization and the performance of recovery, refining-localization-based image authentication and recovery scheme are proposed using iterative compensation mechanism. The authentication algorithm contains the basic authentication and the refining authentication. It fuses the twice authentication results and takes several post-pro- cessing measures based on connected tampered region to obtain the accurate authentication result. According to embedding capacity, the recovery scheme uses iterative compensation mech- anism to recover the tampered pixels. Theoretical analysis and simulation results show that the proposed algorithm can improve the accuracy of tamper localization and the probability of recovering the tampered pixels.
出处
《数据采集与处理》
CSCD
北大核心
2013年第1期51-57,共7页
Journal of Data Acquisition and Processing
基金
国家杰出青年科学基金(61025013)资助项目
国家自然科学基金(61073159)资助项目
中央高校基本科研业务费专项资金(2012JBM042)资助项目
关键词
基本认证
精细认证
连通篡改区域
迭代补偿机制
basic authentication
refining authentication
connected tampered region
iterative compensation mechanism