Premium
TH‐A‐211‐02: Effects of Different Parameters on Deformable Image Registration
Author(s) -
Kim J,
Zhong H
Publication year - 2012
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4736234
Subject(s) - image registration , regularization (linguistics) , computer science , artificial intelligence , metric (unit) , smoothness , software , domain (mathematical analysis) , similarity (geometry) , computer vision , image (mathematics) , process (computing) , algorithm , mathematics , mathematical analysis , operations management , economics , programming language , operating system
Deformable image registration algorithms generally consist of three components, similarity metric, transformation, and optimization. In addition to these components, the optimization also includes one or more regularization terms such as the smoothness of the underlying deformation or restrictions based on the physical properties of organs. It is also common for most of the algorithms to take a multi‐resolution approach to capture large initial deformations as well as to speed up the registration process. Understanding and proper choice of such parameters are critical to a successful registration. The effects of different choices of parameters will be presented using public domain algorithms as well as commercial software. Learning objectives: 1. To understand the basic constituents of deformable image registration algorithms. 2. To understand the consequences of the different parameter choices.