摘要
对不同模态的医学图像进行融合处理,可为临床提供新的诊断信息。针对V．Petrovic提出的灰度图像融合方法,给出了一种改进的多尺度图像融合算法。首先对源图像按照V．Petrovic的方法对源图像的梯度细节图像进行融合/分解并得到源图像的多尺度金字塔表示,然后对金字塔中的两个低通分量加权平均后,再进行反变换,即得灰度融合图像。然后以此灰度融合图像作为两幅源图像的公共部分,利用Toet的伪彩色融合方法,得到最终的彩色融合图像。由于该算法兼容了灰度融合和彩色融合的优点,用不同彩色来显示灰度差,使得融合图像在色彩上更丰富,包含更多的细节。将其应用于CT和MRI图像融合,仿真结果得到了证明。
The fusion of different medicinal images can provide new diagnostic information for clinician. In allusion to V. Petrovic＇ s gray image fusion method, an improved approach was presented in this paper. First we disposed the gradient map of two original images with fusion/decomposition step and obtained a multiresolution pyramid representation of original images. Averaging two low -pass components in pyramid and taking inverse transform, we gained the gray fused image. Afterwards taking this fused image as the common component of two original images and using Toet＇ s false color image fusion approach, the final color fusion image was generated. This method holds the merits of gray - level image fusion and false color image fusion and displays difference of gray degree by different color. So the fusion image has more color variation, more details. This algorithm is applied to the fusion of CT and MRI images, the result testified the above standpoint.
出处
《计算机仿真》
CSCD
2007年第3期202-206,共5页
Computer Simulation
基金
国家自然科学基金重点项目资助(70371066)
关键词
小波变换
图像融合
伪彩色图像融合
Wavelet transform
Image fusion
False color image fusion
