German Cancer Research Center, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany.
Comput Methods Programs Biomed. 2010 Oct;100(1):79-86. doi: 10.1016/j.cmpb.2010.02.008. Epub 2010 Apr 20.
Although non-rigid registration methods are available or under development for many specific problems in medicine, rigid and affine registration is an important task that is often performed for pre-aligning images before using non-rigid registration. In this paper, we present a free and open-source application for rigid and affine image registration, which is designed both for developers and for end-users. The application is based on the Medical Imaging Interaction Toolkit (MITK) and allows for inter-modality and intra-modality rigid 2D-2D and 3D-3D registration of medical images such as CT, MRI, or ultrasound. The framework as well as the application can be easily extended by adding new transforms, metrics and optimizers. Thus, developers of new algorithms are enabled to test and use their algorithms more quickly, spending less work on user interfaces. Additionally, the framework provides the possibility to use image masks to restrict the evaluation of metric values by the optimizer on certain areas of the images.
虽然对于医学中的许多特定问题都有可用的或正在开发的非刚性配准方法，但刚性和仿射配准是一项重要任务，通常在使用非刚性配准之前用于对图像进行预对齐。在本文中，我们提出了一个用于刚性和仿射图像配准的免费开源应用程序，该应用程序既面向开发人员，也面向最终用户。该应用程序基于 Medical Imaging Interaction Toolkit（MITK），可实现医学图像（如 CT、MRI 或超声）的模态间和模态内刚性 2D-2D 和 3D-3D 配准。通过添加新的变换、度量和优化器，可以轻松扩展框架和应用程序。因此，新算法的开发人员可以更快地测试和使用他们的算法，在用户界面上花费更少的工作。此外，该框架还提供了一种可能性，即使用图像掩模来限制优化器在图像的某些区域上评估度量值。
