一种基于零水印的医学图像真伪性认证算法

An Identifying Authentication Algorithm for Medical Images Based on Zero-Watermark

提出一种基于零水印的医学图像真伪性认证算法.该算法对原始医学图像进行分块整数小波变换,在每个分块的低频和高频分别提取4个系数,取其最低有效位与文本水印的对应位相异或,得到2个密钥作为零水印,认证时使用这2个密钥恢复水印信息.算法复杂性低安全性高,不改变原始医学图像.结果表明,该算法能够实现对医学图像的真伪性认证,具有较好的篡改定位能力.

An identifying authentication algorithm for medical images based on zero-watermark is presented. The algorithm splits the original medical image to some blocks to do integral wavelet transform, chooses four low-frequency coefficients and four high-frequency coefficients in every block, and chooses the least significant bits,then dose the XOR operations with the corresponding bit of text watermark. We make the two keys as zero-watermark and use them to extract the text watermark to authenticate the image. The algorithm has less complexity and more security, and it does not modify original medical images. The results of experiments show that this algorithm can be used in identifying authentication for medical images,and it has a strong orientation ability for tampering.