基于GHM多小波与Canny边缘检测的医疗图像隐私保护算法研究

Research on Privacy Protection Algorithm for Medical Images Based on GHM(Geronimo Hardin Massopust)Multi-Wavelet Transform and Canny Edge Detection

应用GHM多小波变换对CT图像按能量进行重要度分解,得出LL2、LH2、HL2和HH2.之后对最强信息子图LL2和最弱信息子图HH2进行强Canny边缘检测,利用检测图像内容边缘较多较细且几乎对医疗诊断无用的特点,通过对无效边缘进行删除和异或运算处理,达到隐藏病患信息的目的.算法对病患图像进行边缘修改与信息嵌入,不仅保护常规可视效果下的隐私,而且可以安全携带病患的个人已经诊断信息.

According to different importance degrees,CT images can be distributed into LL2,LH2,HL2 and HH2by GHM multi-wavelet.The edges of LL2（with the highest important degree）and HH2（with the lowest important degree）are detected by strong Canny method.With many thin edges,which are useless for medical diagnosis,patients＇ information can be hid by deletion and comparison processing of invalid edges.Edges of medical images are modified in order to embed patients＇ information,so the privacy can be protected under the normal visualization and the existing diagnosis information of patients can be carried safely.