基于矢量量化压缩编码的数字水印

提出一种基于矢量量化压缩编码(简称VQ编码)技术的水印策略,在对原图像进行VQ编码后,按码书中码字的相似程度对码字进行划分,根据待嵌入水印图像的大小产生一个随机序列作为密钥,然后根据密钥在压缩数据的特定位置嵌入水印。提出的水印策略,其主要特征在于水印既存在于原图像VQ编码后的压缩数据中,也存在于接收端VQ解码后的图像中。压缩后的数据在数据量上远小于原始数据,所以由它替代原图像携带水印,既节省存储空间,也减小了网络传输时间,特别适用于网络环境下的水印嵌入和提取。更重要的是,这种水印策略具有较好的鲁棒性,能够抵抗诸如裁剪、模糊、JPEG压缩等波形攻击和扭转几何攻击。

A digital watermark technology based on vector quantization compression coding

A watermarking scheme based on the technique of VQ coding is proposed. After VQ coding to the original image, the code word is classified according to the level of similarity in codebook. A random serial generated on the basis of the watermark size is taken as the secret key, and then the watermark is embedded into special positions of the compressed data according to the key. The main feature of the proposed watermarking scheme is that the watermark exists both in VQ coded compressed data of original image and the VQ decoded image of receiving end. The compressed data is much less than the original data in data size, so the network transmission time and storage space can be saved by taking compressed data instead of the original image to carry the watermark. The scheme is very advisable for watermark embedding and detecting in network environment. Further more, the scheme has better robustness that can against image attacking, such as cropping, blurring, JPEG compressing and so on.