根据码书中所有码字之间的最近邻关系，采用最近邻方法标记码书使得所有码字与其最近邻码字的标记值不同。在此基础上，运用I/矩阵运算可在矢量量化压缩图像的每n=2m个码字索引中有选择地最多修改m2」个从而嵌入m比特数据。计算机上的仿真实验结果表明：该方法不仅具有较大的嵌入容量，而且具有较好的隐蔽性，可用于矢量量化压缩图像中的信息隐藏。

According to the nearest relationship, the labeled value of every codeword in the labeled codebook by the nearest neighbor method was different from its nearest codeword, based on which, using I/matrix, as many as m bits data can be embedded by selectively changing at most m2 indexes in every n=2m VQ_compressed indexes. The simulated results by the computer showed that the method used for information hiding in VQ_compressed images has larger capacity and better secrecy.