Incorporating Metric Flows and Sparse Jacobian Transformations in ITK
Universit� catholique de Louvain
| Please use this identifier to cite or link to this publication: http://hdl.handle.net/1926/183 |
Published in The Insight Journal - 2006 January - June.
Submitted by Mathieu Decraene on 12-04-2006.
Various metrics have been proposed in the literature for performing intrinsic automatic image to image registration. Among these measures, mutual information is a very popular one because of its robustness and accuracy for a wide variety of applications. In this paper, we propose a filter for performing non-rigid registration by estimating a dense deformation field derived from the mutual information metric. This filter takes place in the ITK PDE deformable registration design like the Demons algorithm of Thirion. We also show how the concept of metric flow is conceptually linked to the concept of metric derivative for a prior transformation model by the transformation jacobian. We also suggest a sparse implementation of the GetJacobian() method for reducing the computation time of a metric derivative for local transformations models.
Data
3 Files (3Mb)
Code
Automatic Testing Results
by Insight-Journal Dashboard
on Mon Mar 20 04:47:56 2006 for revision #2 
expertise: 5 sensitivity: 4.5 
This project passed all of its tests.Click here for more details.
Go here to access the main testing dashboard.
Automatic Testing Results
by Insight-Journal Dashboard
on Thu Mar 9 12:31:35 2006 for revision #1 
expertise: 5 sensitivity: 5 Reviews
Valuable addition to ITK\\\'s PDE registration framework
by Torsten Rohlfing on 09-22-2006 for revision #2 expertise: 5 sensitivity: 4.5
Summary:
The paper implements extensions to ITK\'s PDE registration framework. An implementation of Mutual Information that is suitable for optical flow-type registrations is introduced. A connection between such algorithms and registration methods with parametric transformation models (e.g., B-splines) is also proposed and implemented.
Hypothesis:
Not applicable.
Evidence:
The authors have applied the newly implemented algorithms to a digital phantom and to actual CT lung images.
Open Science:
Source code and some test data are available and included in the downloadable archive.
Reproducibility:
The \"registration-XX.sh\" scripts referenced in the paper do not seem to be in the downloadable files; it is therefore not clear what parameters the example registrations were run with. Only geometrical phantom images are included in the submission; lung images from Figs. 4/5 are not available.
Use of Open Source Software:
The algorithms described in the paper were implemented within ITK\'s PDE nonrigd registration framework.
Open Source Contributions:
All source code is provided. Some data (digital phantom) is provided with the submission. A URL for additional (lung CT) images is given in the paper.
Code Quality:
The code is reasonably well-written, but not documented completely. In the class interfaces, some doxygen tags semantically refer to multiple class members, which is not appropriate since only the first such member will appear in the output documentation. The implementation parts are not very well documented.
Code formating and indentation does not comply with ITK guidelines.
It seems that there is quite a bit of remaining debug code and diagnostic outputs, e.g., in HermosilloMutualInformationFunction::Initialize().
Applicability to other problems:
The algorithms implemented here should be directly applicable to other nonrigid registration problems, e.g., MR brain image registration etc.
Suggestions for future work:
Requests for additional information from authors:
A more detailed discussion of algorithm parameters would be appropriate, i.e., what are good parameter ranges for the different algorithms and for different applications. A table with settings used for the examples in the paper would also be helpful.
Additional comments:
The word "Jacobian" should probably be capitalized in the paper.
In Eq. (14), the index range of the sum is not fully specified.
In the paragraph between Eq. (15) and Eq. (16), is there an extra "2" in the text?
The paper implements extensions to ITK\'s PDE registration framework. An implementation of Mutual Information that is suitable for optical flow-type registrations is introduced. A connection between such algorithms and registration methods with parametric transformation models (e.g., B-splines) is also proposed and implemented.
Hypothesis:
Not applicable.
Evidence:
The authors have applied the newly implemented algorithms to a digital phantom and to actual CT lung images.
Open Science:
Source code and some test data are available and included in the downloadable archive.
Reproducibility:
The \"registration-XX.sh\" scripts referenced in the paper do not seem to be in the downloadable files; it is therefore not clear what parameters the example registrations were run with. Only geometrical phantom images are included in the submission; lung images from Figs. 4/5 are not available.
Use of Open Source Software:
The algorithms described in the paper were implemented within ITK\'s PDE nonrigd registration framework.
Open Source Contributions:
All source code is provided. Some data (digital phantom) is provided with the submission. A URL for additional (lung CT) images is given in the paper.
Code Quality:
The code is reasonably well-written, but not documented completely. In the class interfaces, some doxygen tags semantically refer to multiple class members, which is not appropriate since only the first such member will appear in the output documentation. The implementation parts are not very well documented.
Code formating and indentation does not comply with ITK guidelines.
It seems that there is quite a bit of remaining debug code and diagnostic outputs, e.g., in HermosilloMutualInformationFunction::Initialize().
Applicability to other problems:
The algorithms implemented here should be directly applicable to other nonrigid registration problems, e.g., MR brain image registration etc.
Suggestions for future work:
Requests for additional information from authors:
A more detailed discussion of algorithm parameters would be appropriate, i.e., what are good parameter ranges for the different algorithms and for different applications. A table with settings used for the examples in the paper would also be helpful.
Additional comments:
The word "Jacobian" should probably be capitalized in the paper.
In Eq. (14), the index range of the sum is not fully specified.
In the paragraph between Eq. (15) and Eq. (16), is there an extra "2" in the text?
Statistics
|
|
| Global rating: |
|
| Review rating: |  [review]
|
| Code rating: | 
|
| Views: | 6355 |
| Downloads: | 960 |
Send a message to the author
Information
|
|
| Paper Id: | 83 |
| Keywords: | Metric flow, Local Jacobian, Non-rigid registration, |
| Toolkit: | ITK |
| Revision: | 2 (03-20-2006) |
| Status: | Open for public review |
| View license
Loading license...
| |
Data
|
|
| Full download: | .zip |
| Paper: | view, .pdf |
Share
|
