利用图像局部视觉特征的自适应盲水印

利用人类视觉特征提出了一种DCT域的自适应盲水印嵌入算法。该算法首先将载体图像划分为图像块序列,运用人类视觉特征将图像块分为三类:平滑块序列、边缘块序列和纹理块序列。对不同类块序列嵌入不同强度的水印信息,增强了水印的不可见性。实验结果表明,该算法具有较强的鲁棒性,同时能保证含水印图像的质量。

Adaptive blind watermark using image local visual characteristics

By using the human vision characteristics,this paper presents an adaptive watermark embedding algorithm based on Discrete Cosine Transform(DCT).Authors split the original image into sequence of blocks,and then classify them into three se- quences:smoothing block sequence,texture block sequence and edge block sequence.According to the classification,the watermark components with different intensity are embedded into different block sequences,so as to enhance the transparency of watermark. Experimental results show that the proposed technique is robust and the quality of the watermarked image is guaranteed.