计算机科学与探索 ›› 2016, Vol. 10 ›› Issue (6): 856-866.DOI: 10.3778/j.issn.1673-9418.1506064

• 人工智能与模式识别 • 上一篇    下一篇

可逆数据隐藏算法在医学图像上的应用

袁陶希,吴昊天+,徐依凌,胡晓渭,刘  渊   

  1. 江南大学 数字媒体学院,江苏 无锡 214122
  • 出版日期:2016-06-01 发布日期:2016-06-07

Reversible Data Hiding Algorithm and Its Application to Medical Images

YUAN Taoxi, WU Haotian+, XU Yiling, HU Xiaowei, LIU Yuan   

  1. School of Digital Media, Jiangnan University, Wuxi, Jiangsu 214122, China
  • Online:2016-06-01 Published:2016-06-07

摘要: 可逆数据隐藏技术是一种新兴的信息隐藏技术,能够在提取嵌入的信息后无失真地还原出原始信息,因此在医学、遥感、军事等领域被广泛使用。提出了一种新的基于直方图平移的可逆数据隐藏算法,解决了已有算法应用在医学图像上出现的严重视觉失真问题,同时实现了病人信息的可逆隐藏和图像的对比度增强。新算法针对医学图像的像素分布特点,改进已有算法的预处理过程,自适应地选择最优预处理区间,并利用最大类间方差法分离背景和前景区域,在信息嵌入的过程中,逐步地增强医学图像感兴趣区域的对比度,之后能够无损地还原出原始图像。仿真实验表明,新算法有效消除了视觉失真,对两类不同的医学图像均有良好的对比度增强效果,信息嵌入率也相当可观。

关键词: 可逆数据隐藏, 直方图平移, 对比度增强, 医学图像

Abstract: Reversible data hiding is a kind of information hiding technique that can exactly recover the original image after data extraction from the image with hidden data. Due to this property, it has been widely used in the areas of medical, remote sensing and military. This paper proposes a reversible data hiding algorithm based on histogram shifting, which can alleviate the visual distortion caused by an existing reversible data hiding algorithm, and reversibly hide the patient information into the contrast-enhanced image. According to the characteristics of pixel distribution in medical images, the pre-processing is improved to adaptively choose the optimal intervals and the method of maximum classes square error is used to separate the background and the foreground. In the information hiding procedure, the contrast of the region of interest can be gradually enhanced. Furthermore, the original image can be exactly recovered when needed. The experimental results on two sets of medical images show that the proposed algorithm can effectively reduce the visual distortions that may be caused by the previous algorithm, while the contrast enhancement effects can be achieved with considerable hiding rates.

Key words: reversible data hiding, histogram shifting, contrast enhancement, medical image