Difference between revisions of "ITK Image Registration with CUDA"

From NAMIC Wiki
Jump to: navigation, search
 
(One intermediate revision by the same user not shown)
Line 4: Line 4:
 
==Objective==
 
==Objective==
 
The objective of this project is to Integrate GPU accelerated rigid and non-rigid registration on CUDA platform into ITK framework.
 
The objective of this project is to Integrate GPU accelerated rigid and non-rigid registration on CUDA platform into ITK framework.
 +
 +
==Framework==
 +
The registration framework is shown in the following figure.
 +
 +
[[Image:itkcudaregistration.jpg|700px]]
 +
 +
Registration framework should not violate the existing itk registration framework, so that user will use itk cuda classes the same way as existing itk classes. itkCudaMemory classes provides the interface for GPU memory the same way as are implemented in vtkCudaMemory classes.
 +
 +
==ITK CUDA classes==
 +
* itkCudaMemory classes
 +
* itkCudaImageToImageMetric and its derivatives
 +
* itkCudaTransform and its derivatives
 +
* itkCudaInterpolateImageFunction and its derivatives

Latest revision as of 10:47, 17 January 2008

Home < ITK Image Registration with CUDA

Description

This page explains the details of ITK registration using CUDA.

Objective

The objective of this project is to Integrate GPU accelerated rigid and non-rigid registration on CUDA platform into ITK framework.

Framework

The registration framework is shown in the following figure.

Itkcudaregistration.jpg

Registration framework should not violate the existing itk registration framework, so that user will use itk cuda classes the same way as existing itk classes. itkCudaMemory classes provides the interface for GPU memory the same way as are implemented in vtkCudaMemory classes.

ITK CUDA classes

  • itkCudaMemory classes
  • itkCudaImageToImageMetric and its derivatives
  • itkCudaTransform and its derivatives
  • itkCudaInterpolateImageFunction and its derivatives