摘要
本文提出了一种基于DCT的图象自适应公开水印算法。该算法结合人类视觉系统屏蔽特性,通过不同强度地调整各8×8DCT块内两个相邻中频系数间的大小关系,以自适应嵌入水印信号,真正实现了盲水印提取,且视觉因子的计算在空域内进行,算法复杂度低。实验结果表明该算法是有效的,在确保水印不可见的同时,对常见图象处理操作和一些几何扭曲有较好的鲁棒性。
An imageadaptive public watermarking technique based on DCT is presented in this paper. A binary watermark is embedded into the original image by modifying the polarity of two neighboring middlefrequency coefficients in the 8×8 DCT block that is pseudorandomly selected by using chaotic sequences, and the watermark is adapted to the image by employing the Human Visual System (HVS) characteristics, the watermark extraction need not resort to the original image. Meanwhile, this algorithm has low complexity because it computes visual factors in the spatial domain. Experimental results prove that the embedded watermark is invisible to human eyes, and it features good robustness under common image processing operations and some geometric distortions.
出处
《计算机工程与科学》
CSCD
2003年第4期1-3,17,共4页
Computer Engineering & Science
基金
国家教育部科研基金(2000)
贵州大学科研基金(2001)
