文件列表:

CodeSubmission (0, 2007-07-15)
CodeSubmission\ApplyDisplacementFieldTransform.cpp (6612, 2007-07-13)
CodeSubmission\ApplySimulatedBreathingTransformation.cxx (7669, 2007-07-13)
CodeSubmission\ApplySimulatedBreathingTransformationWithIntensityVariation.cxx (13004, 2007-07-13)
CodeSubmission\ApplySyntheticTPSTransform.cxx (8086, 2007-07-13)
CodeSubmission\ApplyUniformPeriodicTransform.cxx (4994, 2007-06-26)
CodeSubmission\build (0, 2007-07-15)
CodeSubmission\CMakeLists.txt (3385, 2007-07-13)
CodeSubmission\commonItkTypedefs.h (9567, 2007-04-24)
CodeSubmission\commonTypedefs.h (1106, 2006-09-11)
CodeSubmission\CompareDisplacementFields.cxx (59601, 2007-07-13)
CodeSubmission\CompareWarpedImages.cxx (37918, 2007-04-23)
CodeSubmission\config_template.txt (2123, 2007-07-15)
CodeSubmission\Copyright.txt (1785, 2007-07-13)
CodeSubmission\CTCurvatureRegistration.cxx (1048, 2007-07-15)
CodeSubmission\CTCurvatureRegistration.h (2566, 2006-10-22)
CodeSubmission\CTCurvatureRegistration.txx (6386, 2007-07-15)
CodeSubmission\CTDiffeomorphicDemonsRegistration.cxx (1098, 2007-07-15)
CodeSubmission\CTDiffeomorphicDemonsRegistration.h (2013, 2007-07-03)
CodeSubmission\CTDiffeomorphicDemonsRegistration.txx (19516, 2007-07-15)
CodeSubmission\CTFastBlockMatchingRegistration.cxx (1090, 2007-07-15)
CodeSubmission\CTFastBlockMatchingRegistration.h (1796, 2007-07-03)
CodeSubmission\CTFastBlockMatchingRegistration.txx (4090, 2007-07-15)
CodeSubmission\CTLevelSetMotionRegistration.cxx (1075, 2007-07-15)
CodeSubmission\CTLevelSetMotionRegistration.h (2703, 2006-10-22)
CodeSubmission\CTLevelSetMotionRegistration.txx (7409, 2007-07-15)
CodeSubmission\CTNonlinearRegistrationBase.cxx (1070, 2007-07-15)
CodeSubmission\CTNonlinearRegistrationBase.h (4568, 2007-07-13)
CodeSubmission\CTNonlinearRegistrationBase.txx (14352, 2007-07-13)
CodeSubmission\CTNonlinearRegistrationInterface.cxx (1093, 2007-07-15)
CodeSubmission\CTNonlinearRegistrationInterface.h (6697, 2007-07-13)
CodeSubmission\CTNonlinearRegistrationInterface.txx (11561, 2007-07-13)
CodeSubmission\CTNonlinearRegistrationReturnValue.h (1365, 2007-07-15)
CodeSubmission\CTStandardDemonsRegistration.cxx (1076, 2007-07-15)
CodeSubmission\CTStandardDemonsRegistration.h (2660, 2006-10-22)
CodeSubmission\CTStandardDemonsRegistration.txx (6235, 2007-07-15)
CodeSubmission\CTSymmetricDemonsRegistration.cxx (1079, 2007-07-15)
CodeSubmission\CTSymmetricDemonsRegistration.h (2675, 2006-10-22)
CodeSubmission\CTSymmetricDemonsRegistration.txx (6280, 2007-07-15)
CodeSubmission\DownAndUpsampleUtilities.h (32419, 2007-07-04)
... ...

How to use the nonlinear registration framework? Martin Urschler, urschler@icg.tu-graz.ac.at First you have to compile the whole source code in the directory. Therefore create a build directory and use cmake to set up the makefiles. Please note that you need to turn on FFTWD support to use the curvature registration algorithm, further the FFTW library has to be installed on your system. The source code consists of several types of files, there are the registration algorithms, they are easily distinguishable by their suffix "CT". Second there are the synthetic deformation tools, they start with "Apply". Last, there are the quantitative comparison tools which start with "Compare". The rest of the files are utilities. A later release might provide a better structuring of these modules in directories. The registration algorithms are interfaced by the CTNonlinearRegistrationInterface class. We use this interface to get a single entry point into a shared library that might be compiled from the algorithms. However, note that in this code submission we are not using a shared library but instead compile all the algorithms and the executable tool (from "nonlinearRegistration.cpp") simultaneously. The "nonlinearRegistration" tool requires a configuration file, an example can be found in config_template.txt. There you can also find the names of the possible algorithms and the syntax of the paramter settings. As soon as everything is compiled you have to switch to the python scripts in the PythonEvaluationFramework folder. Use startSyntheticEval.py to set up synthetic experiments. There one has to specify beside some other self-explaining options, the data set to use for the experiment, the synthetic transformations, the algorithms to test and the paths to the tools as well as its names. Also the output folder is important, it will be used as the base directory for date-separated experiments. You can give each experiment a name to put some more information into the subfolder name that will be created. Please also have a look at the README file in the framework folder. We provide some examples for configuration files. The framework works like this. It applies the synthetic transformation to the input data and takes the original and the transformed data for registration. After registration the warped image is compared to the original image and the displacement field is compared to the ground truth field. Everything on the inside of the process is performed automatically. We would of course like to hear the problems that occur during using the framework, since this is research code! Every feedback mail will help to improve the code base.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：