基于最小扩展块合并的无损图像认证算法

Lossless Image Authentication Algorithm with Minimal Expansion

为了在图像分块中嵌入水印信息,现有基于分块的无损图像认证算法需要合并较大的图像子区域,导致篡改定位精度降低.为此,提出基于最小扩展图像块合并的无损图像认证算法.该算法自适应地将图像分成若干认证块,每个认证块由若干细粒度基本块构成,采用递归三等分方式和邻接策略串联基本块,形成基本块链表,以基本块为单位扩展认证块.理论分析表明,该算法能根据图像特征自适应调整认证块大小,优化认证块尺度和邻接性.实验结果显示,采用16×16分块模式能获得较高认证精度;与同类算法相比,该算法认证块划分细密,认证块平均大小降低约19.2%,显著提高了篡改定位精度.

As the existing block-wise lossless image authentication algorithms need to merge a larger image subregion to embed the watermark information and result in the degradation of the precision of locating distortions, a novel lossless image authentication algo- rithm with minimal expansion is proposed. In this algorithm, an image is splitted adaptively into some authentication blocks which are composed of a number of fine-grained basic blocks. A recursive trisection method and adjacent strategies are adopted to connect basic blocks in sequence into a link. When the watermark payload can＇t be embedded into an authentication block, the algorithm tries to merge only one basic block from the basic block link into current authentication block. Theoretical analysis shows that the algorithm can adaptively adjust the size of each authentication block according to the image features to optimize the scale and adjacency of the authentication block. Experimental results show that compared with other algorithms, the authentication block division in the algorithm is finer, and the authentication block size reduced by about 19.2% in average, the algorithm significantly improves the tamper localiza- tion accuracy.