摘要
针对多模态医学影像的成像原理,为了弥补各个模态的医学图像的不足,提出了一种基于非下采样Contourlet变换的医学图像融合算法。首先对源图像进行非下采样Contourlet分解,分别得到低频子带系数和高频子带系数,然后对低频子带系数采用区域能量加权的融合规则,高频子带系数则选取区域标准差比例加权作为融合规则,最后进行非下采样Contourlet逆变换,得到融合图像。通过实验对比表明,该算法明显优于小波(Wavelet)、Contourlet、Wavelet+CS(CS为压缩感知)算法,具有更好的融合性能,清晰度更高,是一种可行、有效的图像融合方法。
For the imaging principle of multi-modal medical image, in order to make up for the shortage of the various modes of medical images, a novel medical image fusion algorithm is proposed based on the nonsubsampled contourlet transform （NSCT）. Firstly, two registered source images are decomposed by the nonsubsampled contourlet transform to obtain the low frequency subband coefficients and high frequency subband coefficients. Secondly, for the low frequency subband coefficients, the fusion principle is based on the weight of local area energy. As for the high frequency subband coefficients, we choose the weight of the area standard deviation ratio as a rule. Finally, the fusion image is obtained by the nonsubsampled contourlet inverse transform. The experimental results show that the proposed method is feasible and effective, and it has better fusion performance and higher definition than the wavelet, contourlet, and wavelet＋＋CS （CS.. compressive sensing） algorithms.
出处
《激光与光电子学进展》
CSCD
北大核心
2013年第11期94-99,共6页
Laser & Optoelectronics Progress
基金
国家自然科学基金(61102008)
教育部重点实验室开放基金(IPIU012011006)
北方民族大学科研项目(2011Y021)
