Image Registration in Medical Imaging: Applications, Methods, and Clinical Evaluation
In this chapter, we describe in detail algorithms that tackle specific clinical problems and that have been extensively validated in close collaboration with radiologists. Our hope is that its contents will be a useful overview of some of the problems tha
- PDF / 1,150,385 Bytes
- 51 Pages / 439.37 x 666.142 pts Page_size
- 16 Downloads / 205 Views
Image Registration in Medical Imaging: Applications, Methods, and Clinical Evaluation Gerardo Hermosillo, Christophe Chefd’hotel, K.-H. Herrmann, Guillaume Bousquet, Luca Bogoni, Kallol Chaudhuri, Dorothee R. Fischer, Christian Geppert, Rolf Janka, Arun Krishnan, Berthold Kiefer, Ines Krumbein, Werner Kaiser, Michael Middleton, Wanmei Ou, J€ urgen R. Reichenbach, Marcos Salganicoff, Melanie Schmitt, Evelyn Wenkel, Susanne Wurdinger, and Li Zhang Abstracts In this chapter, we describe in detail algorithms that tackle specific clinical problems and that have been extensively validated in close collaboration with radiologists. Our hope is that its contents will be a useful overview of some of the problems that present themselves in real clinical practice and of some of the techniques that have proven to fulfill the challenging requirements of these problems in terms of speed and robustness, and that have become properly validated products. Keywords Image registration Motion compensation Multi-resolution approach
11.1
Introduction
Image registration refers to the ability to establish correspondences between locations in two or more images. The need for this ability is ubiquitous in medical imaging. Automatic or semiautomatic methods that bring images into a common reference frame are keystone components of many clinical applications such as analysis of contrast agents in perfusion imaging, lesion localization and comparison in prior studies, atlas based methods for learning normal and abnormal anatomy, analysis of therapy effectiveness on tumors, alignment of the imaging device depending on the region of interest, evaluation of stenosis in arteries and veins, motion analysis of heart walls and indexing methods that group images of similar characteristics. The large number of modalities and protocols that are used in the evaluation of patients nowadays makes it challenging to design automatic methods to put their complementary data into correspondence. At the same time, the images being acquired are of ever increasing resolution and size and the fast-paced
M. Salganicoff (*) Siemens Computer Aided Diagnosis & Therapy, Malvern, PA, USA e-mail: [email protected]
A. El-Baz et al. (eds.), Multi Modality State-of-the-Art Medical Image Segmentation and Registration Methodologies, DOI 10.1007/978-1-4419-8204-9_11, # Springer ScienceþBusiness Media, LLC 2011
263
264
G. Hermosillo et al.
environment created by large volume of procedures and patients, or by emergency situations, create considerable challenges for automatic registration algorithms in terms of performance and reliability. In clinical applications, having access to sophisticated automatic methods for registration may improve the patient outcome and the physician’s effectiveness and efficiency. In summary, registration is a challenging and, in most cases, ill-posed problem that nevertheless is part of everyday clinical problems in need for solutions. Mathematically, the registration problem for three-dimensional scalar images can be stated
Data Loading...
