对原始信号做DCT的算法：Cox 和Piva 等人提出的DCT技术的经典之作。Cox利用随机数发生器产生标准正态序列作为水印信息对图像进行整体DCT变换后，选取除去DC系数之外部分较低频率系数叠加水印信息 Piva则修改整幅图像的中频部分。对原始信号分块后，再作DCT的算法：Hsu和Wu 把图像进行8*8分块，将一个二进制序列作为水印放入DCT的中频区；有些学者则计算整个图像的DCT，把一个实数序列嵌入DCT的中频系数上。选择中频区的好处是一方面尽量减少嵌入信息对图像主观视觉的影响；同时，尽量避免有损压缩对水印信息可能带来的损失。还有一种DCT方法就是把水印信息嵌入到高频系数上，但是采用这种方法，抗压缩性非常差。-done to the original signal DCT algorithm : Cox Piva and others made the DCT technology classic. Cox using random number generator producing standard normal sequence information as a watermark image of the overall DCT, DC coefficient selection removed outside some lower frequency coefficient superposition watermark information Piva will be replaced by images of the entire IF part. The original signal block, before making the DCT algorithm : Wu Hsu and put images 8 * 8 block, a binary sequence as watermarks Add DCT IF; Some scholars calculated whole image of DCT, a real sequence embedded IF DCT coefficients. IF the choice is a good one to minimize the embedded information to the subjective visual images; Meanwhile, try to avoid Compression right watermark information possible losses. There