摘要
提出一种新的图像窜改检测与自恢复算法,不但可以检测定位出图像中任何细微的窜改,而且可以恢复被较大面积窜改的图像数据。首先,在环面自同构映射的基础上采用列偏移方法优化块序列,随后将关联块的组合水印分别嵌入到各映射块内每个像素的三位最低有效比特,最后按照像素尺寸对图像窜改进行逐级检测,并以最小分块为单元来自恢复图像。经实验证明,该算法能有效地进行窜改检测与内容恢复,尤其窜改区域较大时具有显著的效果。
This paper gave a new algorithm for image tamper detect and self-recovery, which could not only detect the location of any minute changes in images but also recover the largely altered image data. Firstly, based on torus automorphism mapping, adopted column sidelong move method to optimize the position of columns. Then embedded the combined watermark of the correlative block to 3 LSB of each pixel in the mapping block respectively. Finally, the image blocks of different sizes could be detected and the smallest block was viewed as the unit for image self-recovery. Experiment results show that the proposed algorithm has significant effect on tamper detection and recovery, especially when the tamper area is very large.
出处
《计算机应用研究》
CSCD
北大核心
2009年第10期3938-3940,3950,共4页
Application Research of Computers
基金
国家科技支撑计划资助项目(2006BAD11A17)
鲁东大学科研资助项目(L20076101)
关键词
数字水印
窜改定位
自恢复
关联块
digital watermarking
tamper locating
self-recovery
correlative block
