【摘要】 提出了一种基于内容自适应的顽健水印算法,以优化水印算法中顽健性和不可见性间的矛盾约束,其主要特点有:构造了小波域熵掩码,并发展成综合人类视觉特性(HVS)模型;给出并分析了基于能量、熵掩码和综合HVS的3种内容自适应测度;用能纠正删除和错误的RA(repeat-accumulate)码实现水印嵌入位置同步;用后验小波域隐马尔可夫模型进行水印检测。实验表明,相对于传统的随机嵌入算法,内容自适应算法能在不降低视觉质量的情况下较大程度地提高水印系统的顽健性。更多还原

【Abstract】 A robust watermarking algorithm based on content-adaptive approach was proposed to optimize the tradeoff between robustness and imperceptibility,which was characterized as follows: the entropy masking proposed by Watson etc.was constructed in wavelet domain,which was then used to constitute the integrated HVS model;three different measures,namely,energy,entropy and integrated HVS,for content-adaptive selection were constructed and investigated;repeat-accumulate(RA) code with erasure and error correction was used to tackle the synchronization issue of the exact positions for watermark embedding;and a posteriori DWT-HMM detector was utilized for watermark detection.Simula-tion results show that the proposed content-adaptive watermarking algorithm,under the same imperceptibility,has con-siderable improvement in robustness performance over the traditional one with stochastic approach.更多还原