摘要
针对传统像素级图像融合方法割裂像素间联系的问题,将医学图像融合与粒度计算相结合,从粒度的角度研究医学图像融合技术,提出基于相容粒度空间的医学图像融合算法。该算法通过将待融合源图像进行小波变换,然后对小波系数构造多层次的相容粒度,最后选择合适的层次进行粒度融合并进行小波逆变换形成最终的融合图像。实验结果表明,该算法在MRI与MRA的图像融合中是有效的。
Focusing on the most traditional image fusion algorithms that split relationship among pixels. Combined with the medical image fusion and the granular computing, research of medical image fusion in granularity, this paper proposed an algorithm for medical image fusion based on tolerance granular space. This algorithm decomposed each source image by wavelet transform, and constructed multi-level tolerance granularity in wavelet coefficients. Finally, selected the appropriate level to fusion image by granularity and through the inverse wavelet transform to form the final fused image. The results of experiment show that this algorithm is effective in MRI and MRA image fusion.
出处
《计算机应用研究》
CSCD
北大核心
2010年第3期1192-1194,共3页
Application Research of Computers
基金
国家自然科学基金资助项目(60775035)