2-D and 3-D Image Registration for Medical Remote Sensing

A. Ardeshir Goshtasby, (2008) 2-D and 3-D Image Registration for Medical Remote Sensing. /E-Books/Mei Sheng - Calculus/.

Full text not available from this repository.


bImage registration is the process of spatially aligning two or more images of a scene. This basic capability is needed in various image analysis applications. The alignment process will determine the correspondence between points in the images, enabling the fusion of information in the images and the determination of scene changes. If identities of objects in one of the images are known, by registering the images, identities of objects and their locations in another image can be determined. Image registration is a critical component of remote sensing, medical, and industrial image analysis systems. This book is intended for image analysis researchers as well as graduate students who are starting research in image analysis. The book provides details of image registration, and each chapter covers a component of image registration or an application of it. Where applicable, implementation strategies are given and related work is summarized. In Chapter 1, the main terminologies used in the book are defined, an example of image registration is given, and image registration steps are named. In Chapter 2, preprocessing of images to facilitate image registration is described. This includes image enhancement and image segmentation. Image enhancement is used to remove noise and blur from images and image segmentation is used to partition images into regions or extract region boundaries or edges for use in feature selection. Chapters 3–5 are considered the main chapters in the book, covering the image registration steps. In Chapter 3, methods and algorithms for detecting points, lines, and regions are described, in Chapter 4, methods and algorithms for determining the correspondence between two sets of features are given, and in Chapter 5, transformation functions that use feature correspondences to determine a mapping function for image alignment are discussed. In Chapter 6 resampling methods are given and in Chapter 7 performance evaluation measures, including accuracy, reliability, robustness, and speed are discussed. Chapters 8–10 cover applications of image registration. Chapter 8 discusses creation of intensity and range image mosaics by registering overlapping areas in the images, and Chapter 9 discusses methods for combining information in two or more registered images into a single highly informative image. In particular, fusion of multi-exposure and multi-focus images is discussed. Finally, Chapter 10 discusses registration of stereo images for depth perception. Camera calibration and correspondence algorithms are discussed in detail and examples are given. Some of the discussions such as stereo depth perception apply to only 2-D images, but many of the topics covered in the book can be applied to both 2-D and 3-D images. Therefore, discussions on 2-D image registration and 3-D image registration continue in parallel. First the 2-D methods and algorithms are described and then their extensions to 3-D are provided. This book represents my own experiences on image registration during the past twenty years. The main objective has been to cover the fundamentals of image registration in detail. Applications of image registration are not discussed in depth. A large number of application papers appear annually in Proc. Computer Vision and Pattern Recognition, Proc. Int’l Conf. Computer Vision, Proc. Int’l Conf. Pattern Recognition, Proc. SPIE Int’l Sym. Medical Imaging, and Proc. Int’l Sym. Remote Sensing of Environment. Image registration papers frequently appear in the following journals: Int’l J. Computer Vision, Computer Vision and Image Understanding, IEEE Trans. Pattern Analysis and Machine Intelligence, IEEE Trans. Medical Imaging, IEEE Trans. Geoscience and Remote Sensing, Image and Vision Computing, and Pattern Recognition. The figures used in the book are available online andmay be obtained by visiting the website http://www.imgfsr.comook.html. The software implementing the methods and algorithms discussed in the book can be obtained by visiting the same site. Any typographical errors or errata found in the book will also be posted on this site. The site also contains other sources of information relating to image registration

Item Type: Article
Uncontrolled Keywords: Gambar, 2-D, 3-D, Medical remote sensing
Depositing User / Editor: Users 1 not found.
Last Modified: 04 May 2012 23:39
URI: http://digilib.uin-suka.ac.id/id/eprint/641

Actions (login required)

View Item View Item