分区域的医学图像高容量无损信息隐藏方法

针对医学图像的分区域典型特征，提出一种基于区域和直方图平移的高容量无损信息隐藏方法。本方法用最大类间距分割法求得原始图像的前景区域，再用聚合多边形逼近和图像拟合法得到其数据嵌入区域。在数据嵌入过程中，提出利用差值直方图循环平移和基于编码的直方图平移方法分别在前景和背景区域嵌入数据，提高了原始直方图平移方法容量和解决了溢出问题。实验结果表明该方法总的嵌入容量可达1 bit/packet以上，并且隐秘图像质量在40dB左右，适用于具有区域特征的质量敏感图像的大容量信息隐藏。

Region-based lossless data hiding with high capacity for medical images

In view of the significant regional features of medical images,a new high capacity lossless data hiding method based on region and histogram shifting was presented.Firstly,the presented method utilized segmentation method of the maximum class separation distance to get the foreground region of medical images,and then used the polymeric polygonal approximation and image fitting algorithm to get the embedding region of region foreground.In the embedding procedure,the difference histogram circular shifting method and the code-based histogram shifting method were presented to embed hidden data in foreground and background region respectively,which improved the limit embedding capacity and resolved the overflow and underflow in traditional histogram shifting method.Experimental results show that the proposed method achieves high embedding capacity with 1 bit/pixel and stego-image quality about 40 dB.The presented method can be applied to the quality of sensitive image's large capacity information hiding with regional feature.