【作者】 张静；

【导师】 李金；

【作者基本信息】 哈尔滨工程大学 ， 生物医学工程， 2016， 硕士

【摘要】 现如今,人们生活水平不断提高,自我保健意识逐渐增强,对自身健康也有了更多的需求。随着医疗事业如日中天的发展,医疗影像设备不断的更新换代,出现了适用于不同设备的二维医学图像。如,CT图像能清晰的显示人身体内各种脏器的解剖结构,PET图像能清晰显示人身体内功能结构。充分利用医学图像融合技术,将“功能显像”与“解剖显像”结合起来,实现优势互补。传统的医学图像融合技术虽然取得了很大的成功,但还是存在许多问题:融合规则没有统一性,针对不同融合对象的算法有差异等等。本文主要针对医学图像中脑部的CT和MRI、MRI和SPECT、MRI和PET三组图像经过多种算法和本文算法进行融合比较。为了建立图像融合的框架,本文分析了医学图像的成像原理和特点,概括性的总结了图像融合的步骤。同时,针对融合图像层次不同的特点将评价标准分成两大类。最后,本文重点分析了基于小波变换的医学图像融合算法和基于Curvelet变换的医学图像融合算法,针对基于小波变换融合后的图像无冗余信息量,但分解过程具有方向局限性,以及基于Curvelet变换的图像融合算法具有多方向性的特点,在小波变换的基础上将低频子带采用加权平均的算法,高频子带采用基于Curvelet变换的区域能量取最大的算法。实验选取10组连续的脑部MRI和SPECT图像以及10组连续的MRI和PET图像进行融合,实验结果通过客观评价标准得出一组平均值。再将本文算法的融合结果与其他算法通过7种客观评价标准进行分析比较。通过分析结果,本文的算法达到了预期效果,经本文算法融合后的图像清晰,集中了源图像的大部分能量,纹理细节反差小,无失真现象,综合评价优于其他算法。在临床医学诊断中,对病灶的定位、观测以及治疗方案的制定有重要的意义。更多还原

【Abstract】 The improved living conditions and enhanced consciousness of Self Health Care of people, make them pay more attention to their heath. Nowadays, medical service is developing rapidly and medical imaging equipment are changing for the better day by day.According to the needs of different medical equipment,there are different types of two-dimensional medical images. For example, CT image can make a video picture of the body, which can clearly display the anatomic structure and then PET images can display function in the human body. Combining the anatomy imaging with function imaging will realize complementary advantages and make the best of Image fusion technology.Though traditional Image fusion technology achieved conspicuous results, many problems still exist: There are no uniform rules of image fusion, there are algorithm difference for different targets etc. This thesis is mainly aiming at 3 groups of brain Image fusion, and these medical images groups are CT and MRI, MRI and SPECT, MRI and PET.This thesis firstly explained the background and significance the research subject,as well as the development status of image fusion technology. And then,the thesis introduced the imaging theory and features of medical image in detail, and generally concluded the procedures of image fusion, and drew a general framework for image fusion. Meanwhile,the thesis made three categories of evaluation criterion,for different levels of image fusion.This paper mainly introduces the medical image fusion algorithm based on wavelet transform and Curvelet transform. Fusion image based on wavelet transform without redundant information in the end, but its limitations is the direction, in the process of decompose,nevertheless,image fusion algorithm based on Curvelet transform has more direction. Therefore, it put forward a new algorithm that low sub-band of decomposition using weighted average fusion algorithm and high sub-band using regional energy maximum algorithm of Curvelet transform on the basis of the wavelet transform. Through the brain images in MRI/SPECT, MRI/PET continuously in ten groups are fusing. It is concluded that a group of average according to evaluation criteria of objective.Finally, comparing the algorithm to image fusion algorithm in common in this paper that through subjective and objective evaluation criteria have proved the quality of the fused images is good in this paper’s algorithm, better than other algorithms, and has great significance in clinical medicine.更多还原