摘要
文中提出了一种新的基于磁性微泡对比剂的超声影像和磁共振影像之间的融合算法.首先利用磁性微泡对比剂对体模进行超声和磁共振显影,并对显影进行边缘分割.运用非采样contourlet变换(nonsubsampled contourlet transform,NSCT)对源图像进行分解,然后对低频子带系数采用自适应动态加权非负矩阵分解(dynamic weighted non-negative matrix factorization,DWNMF)结合分割图像进行融合,然后对各带通方向子带系数采用空间频率激励的脉冲耦合神经网络进行融合,最后利用NSCT逆变换得到融合图像.研究结果表明,利用磁性微泡对比剂并结合上述算法对超声影像和磁共振影像进行融合,可以获得预期的融合效果.
With the help of magnetic microbubble contrast agent, a new fusion scheme between ultrasound image and magnetic resonance image is presented in this paper. Magnetic microbubble contrast agent is used in both ultrasound and NIHI imaging process to enhance and preserve most important information of interest. By using nonsnbsampled contonrlet transform （NSCT）, source images are decomposed. Then, image edge segmentation, combine（t with an self-adaptive algorithm of dynamic weighted non-negative matrix factorization（DWNNIF） is utilized to complete the fusion of low-frequency subband coefficients, the fusion course of bandpass directional subband coefficients can be realized by use of the model of spatial frequcucy-motivated pulse coupled neural networks. Finally, the ultimate fused image can be obtained with inverse NSCT. The experiment based on the above-mentioned schelne demonstrates that ultrasonnd-MRI fusion may be a promising method for obtaining high quality fusion image.
出处
《中国科学：信息科学》
CSCD
2013年第11期1445-1459,共15页
Scientia Sinica(Informationis)
基金
国家重点基础研究发展计划(批准号:2011CB933503)
国家自然科学基金(批准号:60725101
31000453
61179035
61127002)
江苏省基础研究计划(批准号:BK2011036)资助项目
关键词
磁性微泡对比剂
微泡造影剂
融合
动态加权非负矩阵
非采样CONTOURLET变换
magnetic microbubble contrast agent, microbubble contrast agent, fllsion, weighted nonnegative matrix factorization, nonsubsampled contom＇let transform（NSCT）
