本文为布拉格捷克理工大学（作者：Marek Dvorožňák）的硕士论文，共81页。

本文研究了一种可以用于图像预处理的图像配准方法。该方法特别适合于对捕捉关联物体（例如图形）的图像进行配准。本文综述了用于图像配准的基本变形模型，重点研究了保持刚性的模型。还概述了基本的图像配准方法。本文在GIMP的开发版本中实现了一个保持刚性的图像配准工具，并对其实现过程进行了详细的描述。本文对其功能进行了实验，并给出了实验结果。

This thesis deals with image registration method that can be employed for image pre-processing for the purpose of facilitating the process of feature extraction. This method is particularly suitable for registering images capturing articulated objects, for example figures. The thesis includes an overview of basic deformation models employed in image registration, while it focuses on models that preserves rigidity. An overview of basic image registration methods is also included. A tool allowing to perform image registration that preserves rigidity was implemented into development version of GIMP, details of its implementation are described in the thesis. Experiments presenting its functionality were performed and their results are included in this thesis.

1. 图像形变

2. 图像配准

3. 具体实现

4. 实验

更多精彩文章请关注公众号：