基于小波变换的医学图像融合技术的实现

Realization of Medical Image Fusion Based on Wavelet Transform

研究目的:为了对医学图像进行基于小波变换的融合。方法:首先通过学习和研究小波理论的有关知识来研究小波理论在图像融合中的应用目的和实现方法,并借助MATLAB平台,通过小波工具箱实现了图像融合,同时比较了使用这两种方法实现图像融合的效果;然后借助MATLAB中的函数,通过编程实现了图像融合算法。结果:由两幅非同源的医学图像(CT图像和MRI图像)的融合结果可见,两种方法得到的融合图像的效果一致,其对于两幅图像中同一部位相对位置偏移量小的融合效果较好,但对偏移量较大以及存在形变情况的融合则效果不好。结论:通过小波工具箱可以实现一些简单的医学图像融合,但是随着医学图像融合技术的进一步发展和医学图像的复杂度的进一步加深,尤其是对于腹部和胸部的医学图像,则要通过非刚性配准之后才能再进行融合显像,其过程更加复杂。

Purpose： In this paper, we accomplish image fusing of CT and MRI images using wavelet. Methods： As wavelet is of comfortable effect in image fusing, more and more researchers are working on it. And, in Matlab, there is a toolbox called Wavelet Toolbox 3.0, with which we can finish some image fusing simply. With some drawbacks, the toolbox works not so well with some kinds of image pairs, such as a picture of ct and a picture of pet, if the pet image is gray image instead of a counterfeit color image. Then the resulted image if of bad quality with low-lying effect. In order to achieve a good result of image fusing we should think about many aspects of it and realize them by program. Result： We strive to fulfill the program and compare the result to what We have got by toolbox. And the effect is comfortable. Conclusion： By wavelet toolbox, some simple medical image fusing can be accomplished. But with the development of technique of image fusing and the deepening of complexity of medical images, the process of image fusing would be more complex especially those images rooted from abdomen and chest. So more methods must be called for rather than toolboxes.