摘要
水印的不可感知性是评价水印算法性能优劣的重要指标之一,传统图像质量评价模型PSNR不能很好地评价水印化图像的质量。本文通过分析彩色图像中背景亮度、纹理、频率以及颜色对人类视觉的掩蔽,提出了一个基于DWT的彩色图像视觉掩蔽强度计算方法,并据此为彩色图像中各像素设置视觉质量掩蔽权值,得到了一个彩色图像水印不可感知性评价模型。实验结果表明新模型优于传统模型PSNR,其评价结果更接近人的视觉感觉。
The imperceptibility of the watermark is one of the most important aspects for the performance of the watermarking algorithrru Conventional measure of image quality PSNR can't evaluate the quality of the watermarked image well. The different background luminance, texture, frequency and the color of the image mask the human eyes differently. This characteristic can be used to improve the conventional measure PSNR. By calculating the masking threshold of the background luminance, the texture, frequency and the color of pixeI, the masking-weight for each pixel was calculated and an imperceptibility measure with these weights was obtained for the watermark in the watermarked color image, The conventional measure and the new one were compared through different experiments. The result judged by the new measure is more close to the result judged by human eyes than the conventional measure.
出处
《计算机科学》
CSCD
北大核心
2008年第10期208-211,共4页
Computer Science
基金
国家自然科学基金(70371052)
山东省优秀中青年科学家奖励基金(2006BS01021)
山东科技大学科学发展基金(05g018)
关键词
数字水印
不可感知性评价
HVS
DWT
Digital watermarking,Imperceptibility evaluation, Human vision system,Discrete wavelet transform