中文摘要：

为了减小图像自嵌入水印的长度,降低水印对载体图像质量的影响,同时为消除基于分块的图像认证算法中图像块之间的独立性,提出了一种变容量的自嵌入易碎水印算法：首先对图像进行2＆#215;2的分块,根据各分块灰度均值生成原始图像的均值图像;根据均值图像各像素之间的相关性进行游程编码;将编码信息作为水印嵌入在图像像素的最低两位中;篡改检测时首先使用字符串匹配的思想进行图像块和水印之间的匹配,对于未匹配成功的块,使用分组的方式进行再次匹配,以完成认证和恢复.该算法进行一次的水印嵌入,同时用于篡改检测与恢复,水印长度依赖于均值图像相邻像素之间的相关性,可有效缩短水印长度,减小对载体图像质量的影响.理论分析和实验仿真表明,算法在不可见性、篡改定位和恢复、抗拼贴攻击、漏检率等方面具有较好的效果.

英文摘要：

The existing fragile watermarking algorithms based on blocks can＇t drastically avoid the independence between the blocks. A novel variable-payload self-embedding fragile watermarking algorithm is proposed. Firstly, the original image is divided into blocks with the size of 2 × 2 and the average value of each block is calculated. Then a new image called average-image can be composed of the obtained average values. The watermark is not generated for every signal block, but encoded statistically according to the correlation between the pixels of averageqmage. The watermark is embedded into 2-LSB （least significant bit） of the original image. As a result, the length of the watermark is variable due to the correlation of the image. That is to say, for a texture complex image the watermark is longer, but for a smooth image the watermark is shorter. And moreover, the independence of the blocks can be broken, so it has a better ability to resist attacks. As the authentication, string matching is used to match the blocks and the watermark. For the blocks which are not matched, group matching is used to re-match. The theoretic analysis and the simulation results show that the proposed algorithm can not only thwart the collage attack and locate tampered blocks accurately, but also has better non-visibility.