摘要
将具有平移不变性和良好的方向分析能力的双树离散小波引入医学图像融合的方法中,提出了高频子带采用基于跨尺度的邻域空间频率的融合策略,最高层的低频子带采用基于改进的邻域熵的融合策略进行融合处理的融合算法.实验结果表明该方法能够更有效的实现融合,有着良好的融合效果.
Shift invariant and directionally selective dual tree complex wavelet transform were introduced, and across-band-neigh- borhood-space-frequency and improved-neighborhood-entropy were defined for high-pass and low-pass sub-bands respectively. The results demonstrate that this method is more efficient in image fusion than real wavelets.
出处
《山东大学学报(工学版)》
CAS
2008年第1期32-35,42,共5页
Journal of Shandong University(Engineering Science)