一种高容量盲检测图像指纹算法

A high capacity image fingerprinting scheme with blind detection

针对共谋攻击,提出了一种高容量的空域盲检测图像指纹算法。用抗共谋攻击码(ACC)调制正交基向量生成二值指纹,根据指纹状态对图像灰度值量化嵌入指纹。追踪共谋者时,首先根据待检图像像素灰度所属区间提取指纹,然后计算其与正交基向量的内积得到一新向量,对其用软阈值方式进行处理,最后追踪共谋者。指纹嵌入提取采用量化方式,是一种盲检测算法。实验结果表明,该算法不但提高了指纹容量,而且比现有的盲检测方法更有效。

In order to resist the collusion attack,a high-capacity and blind-detection fingerprinting scheme for image is proposed in spatial domain.The orthogonal basic vectors are modulated with anti-collusion codes to generate binary fingerprints,and all pixels of the original image are quantized according to fingerprint states for fingerprint embedding.During colluder tracing,the fingerprints are extracted by judging pixels′ gray levels of the tested image at first,and then a new vector is derived from the inner product between the extracted fingerprints and the orthogonal basic vectors,which is processed with soft threshold.The colluders are traced at last.The quantization is employed in fingerprint embedding and extracting,which is a blind detection.Experimental results show that the scheme not only improves the capacity seriously,but also is more efficient than existing blind detection algorithms.