基于形态学金字塔的医学图像融合技术

Fusion Technology of Medical Image Based on Morphological Pyramid

针对像素级医学图像的融合问题，设计一种形态学金字塔的医学图像融合技术。先对医学图像进行形态学的平滑处理，再进行形态学金字塔分解，最后对形态学金字塔进行反变换以获取最终的融合图像。从平均梯度、联合熵、相关系统对图像融合质量做客观评价，并以WindowsXP的操作系统为实验环境进行融合实验。从融合效果来看，丈中技术能将CT图像和MRI图像融合成一幅信息量饱满的单一图像。实验结果表明：该技术与同类其他方法比较，平均梯度能提高10．30％～36．76％，联合熵能提高5．04％～36．83％，相关系数能提高0．94％～31．52％。

Focusing on the fusion problem of the medical images for pixel level, the medical image fusion technology based on morphological pyramid was designed. First the medical images were processed morphological smooth, second the source images were decomposed using morphological pyramid, and finally the fused image was achieved through inverse morphological pyramid transform. From the fusion effect, the technology can fuse into a single image of full information from CT image and MRI image. Experimental results and comparisons with three traditional methods demonstrate that the designed technology is superior with 10.30% to 36.76% improvement of average gradient, 5.04% to 36.83% enhancement in union entropy and 0.94% to 31.52% improvement of correlation coefficient.