摘要
在抵抗JPEG压缩的同时保持较好的篡改检测能力是用于图像内容认证的半易损水印的关键问题.提出一种基于格的半易损水印算法,将图像的DCT变换系数分组量化为高维空间中的格点,通过比较格点坐标与DCT系数之间的关系自适应地在格点中嵌入水印信息.实验结果表明,与同类算法相比该算法具有较好的性能,含水印图像的峰值信噪比超过49dB,在抵抗质量因子为50的JPEG压缩的情况下仍具有较好的篡改检测和定位能力.
The key point of semi-fragile watermark for image content authentication is resisting to JPEG compression while maintaining high detection ability to tamper. In this paper, a semi-fragile watermark based on lattice is proposed. First, the DCT coefficients of original image are grouped and quantized into lattice point in high-dimension space. Then, information bits are adaptively embedded into lattice according to comparing results between the DCT coefficient and coordinate of high-dimension lattice point. The Experimental results demonstrated that the proposed algorithm has the advantages than the others, PSNR of the watermarked image is larger than 49dB, and can effectively detect and localize the tampered area at the same time resist against JPEG compression with quality factor as low as 50.
出处
《小型微型计算机系统》
CSCD
北大核心
2009年第2期327-331,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60402008
60533100)资助
上海市教委科技发展基金项目(05AZ71)资助
上海市教委优秀青年教师科研专项基金项目资助
关键词
半易损水印
离散余弦变换
格矢量量化
semi-fragile watermark
discrete cosine transform
lattice vector quantization
