【信息技术】【2014.05】医学图像的配准技术

本文为印度Rourkela国立技术研究所（作者：Sangeeta Sahu）的硕士论文，共58页。

图像配准是许多实时图像处理应用中的首要步骤。图像配准是将两幅或两幅以上的图像合并到一个坐标系中进行后续分析，有时也被称为图像对齐。它广泛应用于遥感、医学成像、多传感器融合目标识别、利用卫星图像监测某一特定土地的利用情况、从不同医学模式获得的图像对准等疾病诊断，这是图像融合和图像拼接领域的重要一步。

本文提出了一种图像配准方法，图像配准方法可以分为两类。一种是基于图像灰度值的强度方法，另一种是基于图像特征的强度方法，通过获取图像中的点、线、面等特征或标志来实现。在图像中可以很容易地检测出边缘，因此，利用这些边缘可以得到一些特征，从而实现基于特征的配准。但是，基于特征的配准具有一定的局限性和优势。该方法采用基于特征的配准技术得到一幅粗配准图像，并将其作为基于强度的配准技术的输入，从而得到一个良好的配准结果。它有助于减少基于强度的技术的局限性。也就是说，配准所需的时间更少。为了实现这一任务，选择互信息作为相似度参数。互信息（MI）是一种广泛应用的相似性度量方法。为了提高相似性度量的鲁棒性，将空间信息与标准化互信息（NMI）相结合。将MI与梯度项相乘，从而将空间信息集成到互信息中作为相似性度量。如果采样分辨率较低，则对配准功能的影响较小。它包含正确的全局极大值，有时在互信息情况下可能无法找到。为了达到优化的目的，采用了快速收敛的粒子群优化技术（FCPSO）。在该优化方法中，通过将所有粒子的新变量、粒子平均维数（PMD）添加到现有位置和速度方程中，来平衡单粒子位置的多样性，该方法通过减少优化迭代次数来缩短收敛时间。

Image registration is a primary step inmany real time image processing applications. Registration of images is thebringing of two or more images into a single coordinate system for itssubsequent analysis. It is sometimes called image alignment. It is widely usedin remote sensing, medical imaging, target recognition using multi-sensorfusion, monitoring of usage of a particular land using satellite images, imagesalignment obtained from different medical modalities for diagnosis of diseases.It is an important step in the field of image fusion and image mosaicing. Inthis research work, approaches for image registration are proposed. The imageregistration methods can be grouped into two classes. One is intensity basedmethod which is based on gray values of the pair of images and the second oneis based on image feature which is done by obtaining some features or landmarksin the images like points, lines or surfaces. Edges in the images can bedetected very easily in the images. Thus, using these edges some features canbe obtained by which we can accomplish feature based registration. But, featurebased registration has some limitations as well as advantages. The proposed methodemploys feature based registration technique to obtain a coarsely registeredimage which can be given as input to intensity based registration technique toget a fine registration result. It helps to reduce the limitations of intensitybased technique. i.e. it takes less time for registration. To achieve thistask, the mutual information is selected as similarity parameter. Mutualinformation (MI) is used widely as a similarity measure for registration. Inorder to improve the robustness of this similarity measure, spatial informationis combined with normalized mutual information(NMI). MI is multiplied with agradient term to integrate spatial information to mutual information and thisis taken as similarity measure. The registration function is less affected ifsampling resolution is low. It contains correct global maxima which aresometimes not found in case of mutual information. For optimization purpose,Fast Convergence Particle Swarm Optimization technique (FCPSO) is used. In thisoptimization method, the diversity of position of single particle is balancedby adding a new variable, particle mean dimension (pmd) of all particles to theexisting position and velocity equation. It reduces the convergence time byreducing the number of iterations for optimization.

1. 引言
2. 文献回顾
3. 基于轮廓信息和互信息的医学图像配准
4. 基于强度的医学图像刚性配准
5. 结论

更多精彩文章请关注公众号：