自适应量化索引调制算法(AQIM)通过Watson视觉模型来计算量化步长,使得量化步长可以随着待量化系数自适应变化,从而获取了比传统量化索引调制算法(QIM)更好的不可见性和鲁棒性.但即使在没有干扰的情况下,该算法在检测时计算的量化步长与嵌入时计算的量化步长也不一致,这就导致水印不能够完整可靠地提取.本文在分析该算法存在问题的基础上,提出了一种改进的迭代AQIM水印方案,仿真实验表明该方案比原算法具有更好的性能.
Adaptive Quantization Index Modulation(AQIM) calculates the quantization step based on the Watson's visual model,in which the quantization step varies adaptively on the quantized coefficient.So AQIM is more reliable and robust than the traditional Quantization Index Modulation(QIM).But even if there is no noise,the embedding step in watermarking is different from the one in detecting,which results in that the watermark can't be integrally extracted.In the paper,an iterative AQIM is advanced to overcome the defect and has better result than AQIM.