Files in this item



application/pdf3199198.pdf (4MB)Restricted to U of Illinois
(no description provided)PDF


Title:Digitizing a Three-Dimensional Brain Atlas: Image Sequence Alignment and Volumetric Encoding
Author(s):Zhao, Rongkai
Doctoral Committee Chair(s):Belford, Geneva G.
Department / Program:Computer Science
Discipline:Computer Science
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Computer Science
Abstract:A 3D brain digital atlas is an important tool for neuroscience research. Many different imaging technologies such as MRI, CT, PET, microscopic imaging, and wet section photography, etc., are available. Among the technologies, cryosectioning followed by wet section photography can yield the highest resolution image of a full brain section. However, the images in the raw image sequence are not aligned with each other and therefore must be spatially registered. This dissertation describes several new methods employed by the atlas construction process. The most important components of the image registration process are the objective function and the optimization strategy. Pairwise image registration is inappropriate due to the lack of consideration of global coherence. A novel objective function called minimum entropy of bad prediction (MEBP) is proposed. MEBP is based on information theory and can be used for multi-modal image registration as well as image sequence alignment (ISA). In ISA, MEBP concurrently take multiple images into consideration and therefore can yield better alignment result. The optimization algorithm is a new hybrid method composed of density-based clustering algorithm, multi-resolution method and simplex method. This new method is less data-specific and more suitable for semi-automatic or automatic image registration. The image post-processing and volume compression are other important components in atlas construction. The commonly existed Swiss-cheese type image defect is treated with wavelet-based method and long range correlation. An octree variant, scalable hyperspace file (SHSF), was developed to encode the volumetric data set. SHSF can facilitate volume accessing so that efficient virtual brain slicer and surface viewer are possible to implement. Since a high-resolution brain digital atlas can occupy hundreds of megabytes to tens of gigabytes storage space, in order to support applications on the atlas, volume compression algorithms are necessary. Three compression algorithms for different purposes are proposed. They include lossless method, near-lossless method and progressive lossy-to-lossless method. The progressive method is the most versatile and generalized coder. The various topics covered in this dissertation are compiled into a coherent whole with the digital brain atlas as the central scene.
Issue Date:2005
Description:149 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2005.
Other Identifier(s):(MiAaPQ)AAI3199198
Date Available in IDEALS:2015-09-25
Date Deposited:2005

This item appears in the following Collection(s)

Item Statistics