摘要
提出了一种根据像素的空间邻域特性在二值图像中嵌入数字水印的算法.由邻域内的平滑度和连通性计算像素的可非度,从而确定各种邻域组合的可嵌入优先级.为了均衡可嵌入水印像素的空间分布,对原始图像进行随机排列和分块后,在每一块中嵌入相同数量的水印.并根据随机排列的统计特性,自适应地确定可嵌入水印像素,以尽量减少对图像质量的损害.实验结果表明,嵌入的水印不易被察觉,而且是脆弱的,可以作为检测图像是否被改动过的一种有效手段.
This paper proposed a digital watermarking algorithm which can be applied to binary images. Based on the smoothness and connectivity within 3×3 neighborhood, we compute each pixel's degree that can be negated so as to obtain the priority of embeddability for each 3×3 neighborhood pattern. In order to average the spatial distribution of embeddable pixels, by permuting the original image randomly and dividing it into blocks, we embed the same quantity of watermark into every block. According to the statistics of random permutation, the embeddable pixels can be determined adaptively to minimize the degradation of the image quality. The experimental results show that the embedded watermark is invisible. And it is a fragile one which can be used to detect whether the watermarked image has been tampered.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2004年第9期1509-1514,共6页
Journal of Shanghai Jiaotong University
关键词
数字水印
二值图像
数据隐藏
认证
digital watermark
binary image
data hiding
authentication