Difference between revisions of "2010 Winter Project Week RegistrationCaseLibrary"

From NAMIC Wiki
Jump to: navigation, search
 
(60 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 
__NOTOC__
 
__NOTOC__
 
<gallery>
 
<gallery>
Image:PW-SLC2010.png|[[2010_Winter_Project_Week#Projects|Projects List]]
+
Image:RegistrationCaseLibrary_snapshot.png‎|Example Snapshot of Registration Case Library Entry]]
Image:genuFAp.jpg|Scatter plot of the original FA data through the genu of the corpus callosum of a normal brain.
+
Image:Registration_ExamplePresetMenu.png|Example Preset Menu. Presets will be a new feature implemented into the RegisterImage module to organize, share & communicate regisration input parameters tailored to a specific scenario.  
Image:genuFA.jpg|Regression of FA data; solid line represents the mean and dotted lines the standard deviation.
+
Image:SRegVar_Roll1.png|Example of Registration Parameter Exploration: registering a FLAIR  brain to T1 at 1,2,5,10% sampling rates. Such explorations will be part of finding the optimal registration for a given scenario.
 
</gallery>
 
</gallery>
  
Line 16: Line 16:
  
 
<h3>Objective</h3>
 
<h3>Objective</h3>
This work is part of a 2-year effort to supplement the 3DSlicer Registration Module with extensive set of documentation, tutorials and use case scenarios. We seek to build a comprehensive library of use cases for every form of registration challenge using 3DSlicer. Each library entry will contain a dataset along with Registration Parameter Presets and a guided tutorial on how to successfully register the images.
+
This work is part of a 2-year effort to supplement the 3DSlicer Registration Module with extensive set of documentation, tutorials and use case scenarios. We seek to build a comprehensive '''[[Projects:RegistrationDocumentation:UseCaseInventory|library of use cases ]]''' for every form of registration challenge using 3DSlicer. Each library entry will contain a dataset along with Registration Parameter Presets and a guided tutorial on how to successfully register the images.  
 
 
 
 
  
  
  
 +
We also  have a [[AHM2010:RegistrationBreakout|''Registration Breakout Session'']] on Tuesday 1-2pm (Amethyst Room). If registration is on your processing list, please drop in, have a look at what's new, and share your thoughts.
  
 
</div>
 
</div>
Line 29: Line 28:
 
<h3>Approach, Plan</h3>
 
<h3>Approach, Plan</h3>
  
The use case library relies mainly on user contributions for content. There is an [[Projects:RegistrationDocumentation:UseCaseAdvertisement|'''official call for datasets''']]  to all slicer users to consider a unique time-limited service for any dataset/registration problem that is not yet in the library. Users who agree to share one '''anonymized''' dataset for the library will have that dataset registered by us and the solution and strategies posted.  [[Projects:RegistrationDocumentation:UseCaseExample1|Look here for an example of a library entry.]]
+
The use case library relies mainly on user contributions for content. There is an [[Projects:RegistrationDocumentation:UseCaseAdvertisement|'''official call for datasets''']]  to all slicer users to consider a unique time-limited service for any dataset/registration problem that is not yet in the library. Users who agree to share one '''anonymized''' dataset for the library will have that dataset registered by us and the solution and strategies posted.  [[Projects:RegistrationDocumentation:UseCaseExample1b|Look here for an example of a library entry.]]
 
+
<br>
Our plan for the project week is to first  get a survey of all projects using registration. We hope to catch most of these projects from reviewing these project description pages, so it is important that if you use registration in your procedure, please make sure to mention exclusively.  We will collect as many example cases as possible, and select a few focus cases with particularly vexing registration challenges and begin to develop the appropriate tailored registration strategies.  <bar>,...
+
Our plan for the project week is to first  get a survey of all projects using registration. We hope to catch most of these projects from reviewing these project description pages, so it is important that if you use registration in your procedure, please make sure to mention explicitly.  We will collect as many example cases as possible, and select a few focus cases with particularly vexing registration challenges and begin to develop the appropriate tailored registration strategies.   
  
 
</div>
 
</div>
Line 38: Line 37:
  
 
<h3>Progress</h3>
 
<h3>Progress</h3>
*List of contacts for new use cases goes here
+
*see below for list of projects/contacts for new use cases: [[2010_Winter_Project_Week_RegistrationCaseLibrary#Target Projects]]
*Notes re. user insights/comments/wishes on registration go here
+
*consider a '''priority matrix''' for each case that lists the requirements for robustness, precision, speed. E.g. change tracking applications rank high in precision but low in the other two, IGT applications rank high in speed and robustness etc.
 
+
*'''modularity issue''': (entropy) many implement their own registration module ->cannot benefit from synergies/parallel dev elopment outside ITK
 +
*'''MIDAS''': discussed with Steve & Casey the use of MIDAS server to host the case library; this would enable additional use of library cases for automated testing (CTest)
 +
*the '''manual alignment''' aka ''transforms module'' is well used and important to the user community. The rotation history slider and direct image dragging and rotating are frequent requests. Also addition of scale (& shear)
 +
*the ''Fiducial Bspline'' module is back. After being removed in consideration there are now users that would receive great value added from this. Esp. the category of ''Registration Repair'', i.e. small manual corrections to an (automated) calculated Xform.
  
 
</div>
 
</div>
Line 46: Line 48:
  
 
<div style="width: 97%; float: left;">
 
<div style="width: 97%; float: left;">
 +
=== Progress: Target Projects ===
 +
The projects below have been tentatively identified as having registration ties. We will seek discussion with these projects during the week. If your project is listed here, I will try to discuss the registration aspects at some point during the week.  Type and length of discussion will vary greatly. Text next to each project below identifies the registration link.
 +
*'''Segmentation''':
 +
**[[2010_Winter_Project_Week_SegmentationWizard|High Level Wizard for Segmentation of Images]] ('''Mark Scully'''): uses register images to align atlas; functionality and compatibility considerations for RegisterImages redesign (both algorithm & GUI)
 +
**[[2010_Winter_Project_Week_LongitudinalLupusAnalyses|Longitudinal Analyses of Lesions in Lupus]] (Mark Scully):
 +
***uses ChangeTracker as current prototype, although input data differs: T1,T2 +FLAIR are standard channels, currently T1 is used for registration. As with all change tracking and intra-subject registration:
 +
***precision: ***, robustness: **, speed: *
 +
***library addition: yes
 +
**[[2010_Winter_Project_Week_Musco_Skeletal_Segmentation | Rapid Segmentation of Knee Structures for Simulation]] (Harish Doddi, Saikat Pal, Luis Ibanez, Scott Delp)
 +
***excellent registration challenge, requires multiple approaches: masking, fiducials, surfaces
 +
***Library addition: [[Media:RegLib_DATA_knee.zip.zip‎]]
 +
**[[2010_Winter_Project_Week_ProstateSeg|Prostate segmentation using shape-based method]] (Yi Gao):
 +
***Does involve co-registration of an atlas
 +
***uses ITK directly: criteria: faster & robustness over greater rotational range
 +
**[[2010_Winter_Project_Week_SegmentationEpicardialWall|Epicardial Wall Segmentation]] (Behnood Gholami)
 +
***employs its own shape-registration scheme, uses ITK registration algorithm, does not go through slicer interface
 +
***N/A
 +
**[[2010_Winter_Project_Week_TBISegmentation|Segmentation of TBI (Traumatic Brain Injury) Subjects from Multimodal MRI]] ('''Marcel Prastawa'''):
 +
***includes registration in new (ABC) module: category: intra-subject multi-contrast reg. Currrently also uses its own registration algorithm.
 +
***library: UCLA data, to be determined
 +
**[[2010_Winter_Project_Week_Cardiac_Ablation_Scar_Segmentation|Cadiac Ablation Scar Segmentation]] (Michal Depa):
 +
***employs a label fusion segmentation algorithm which first registered all of the training images to the test image
 +
***uses insight journal log domain diffeomorphic daemons
 +
***dataset: in deliberation
 +
*'''Registration'''
 +
**[[2010_Winter_Project_Week_ThalamicNucleiAtlas | Fusion of Anatomy,MRI and Electrophysiology in Parkinson's]]  (Andrzej Przybyszewski): 
 +
***co-registration of subject MRI with anatomical atlas; very small structure target, no intrinsic contrast in MRI
 +
***involves moving image with drastically clipped FOV, acquisition parameters TBD
 +
***electrophysiology data avail. as second/validation channel
 +
***final atlas needs to move into patient-specific stereotactic space, there is no intra-operative scan.
 +
**[[Tissue_Dependent_Registration|Registration with Varying Elastic Parameters]] (Peter Risholm)
 +
**COPD Gene Collaboration (Raul San Jose)
 +
***lung CT datasets for group registration
 +
***library dataset: in deliberation
 +
*'''IGT'''
 +
** [[2010_Winter_Project_Week_MRI_Guided_Robotic_Prostate_Intervention| MRI-guided Robotic Prostate Intervention]] (Andras Lasso and '''Junichi Tokuda'''): Transperineal device: includes registration, where accuracy should be improved; currently the registration is based on a single-slice image; it could possible to improve the accuracy by increasing the number of slices used for the registration.
 +
*'''Radiotherapy'''
 +
** [[Adaptive Radiotherapy for Head, Neck, and Thorax]] (Ivan Kolesov): Once segmentation is available, the goal is to register patient scans to account for anatomical changes between visits
 +
*'''Analysis'''
 +
**[[2010_Winter_Project_Week_VervetMRILongitudinalAnalysis|Vervet MRI Longitudinal Analysis]] (Andriy Fedorov):
 +
***Register follow-up to baseline: non-human data
 +
***solved with avail. registration tools
 +
***dataset: yes
 +
*'''Informatics'''
 +
**'''XNAT''' : discussion with XNAT group re. feasibility of using XNAT central as repository for use-case library.
 +
* '''Diffusion'''
 +
**[[ 2010_Winter_Project_Week_WM_ATLAS|Atlas-Based White Matter Segmentation]] (Lauren O'Donnell)
 +
*'''NA-MIC Kit Internals'''
 +
**[[2010_Winter_Project_Week_SPECTRE_3DSlicer_Integration|Integration of SPECTRE Java module into 3D Slicer]] (Nicole Aucoin): SPECTRE: skull stripping is an important element for registration preprocessing, documentation and possibly GUI will link to this module.
 +
 +
=== Progress: Notes ===
 +
*consider a ''priority matrix'' for each case that lists the requirements for robustness, precision, speed. E.g. change tracking applications rank high in precision but low in the other two, IGT applications rank high in speed and robustness etc.
 +
*'''MIDAS''': discussed with Steve & Casey the use of MIDAS server to host the case library; this would enable additional use of library cases for automated testing (CTest)
 +
*the ''manual alignment'' aka ''transforms module'' is well used and important to the user community. The rotation history slider and direct image dragging and rotating are frequent requests. Also addition of scale (& shear)
 +
*the ''Fiducial Bspline'' module is back. After being removed in consideration there are now users that would receive great value added from this. Esp. the category of ''Registration Repair'', i.e. small manual corrections to an (automated) calculated Xform.
 +
*XNAT Desktop and Slicer Browser: candidate for hosting Case Library
  
 
==References==
 
==References==
 +
*[[2010_Winter_Project_Week#Projects|back to AHM_2010 project list]]
 
*[[Projects:RegistrationDocumentation|Link to the RegistrationDocumentation Project Wiki]]
 
*[[Projects:RegistrationDocumentation|Link to the RegistrationDocumentation Project Wiki]]
 
*[[Projects:RegistrationImprovement|Link to Kitware Registration Developer Efforts]]
 
*[[Projects:RegistrationImprovement|Link to Kitware Registration Developer Efforts]]
  
 
</div>
 
</div>

Latest revision as of 21:57, 7 January 2010

Home < 2010 Winter Project Week RegistrationCaseLibrary

The 3DSlicer Registration Case Library Project

Key Investigators

  • BWH: Dominik Meier, Ron Kikinis
  • Kitware: Casey Goodlett

Objective

This work is part of a 2-year effort to supplement the 3DSlicer Registration Module with extensive set of documentation, tutorials and use case scenarios. We seek to build a comprehensive library of use cases for every form of registration challenge using 3DSlicer. Each library entry will contain a dataset along with Registration Parameter Presets and a guided tutorial on how to successfully register the images.


We also have a Registration Breakout Session on Tuesday 1-2pm (Amethyst Room). If registration is on your processing list, please drop in, have a look at what's new, and share your thoughts.

Approach, Plan

The use case library relies mainly on user contributions for content. There is an official call for datasets to all slicer users to consider a unique time-limited service for any dataset/registration problem that is not yet in the library. Users who agree to share one anonymized dataset for the library will have that dataset registered by us and the solution and strategies posted. Look here for an example of a library entry.
Our plan for the project week is to first get a survey of all projects using registration. We hope to catch most of these projects from reviewing these project description pages, so it is important that if you use registration in your procedure, please make sure to mention explicitly. We will collect as many example cases as possible, and select a few focus cases with particularly vexing registration challenges and begin to develop the appropriate tailored registration strategies.

Progress

  • see below for list of projects/contacts for new use cases: 2010_Winter_Project_Week_RegistrationCaseLibrary#Target Projects
  • consider a priority matrix for each case that lists the requirements for robustness, precision, speed. E.g. change tracking applications rank high in precision but low in the other two, IGT applications rank high in speed and robustness etc.
  • modularity issue: (entropy) many implement their own registration module ->cannot benefit from synergies/parallel dev elopment outside ITK
  • MIDAS: discussed with Steve & Casey the use of MIDAS server to host the case library; this would enable additional use of library cases for automated testing (CTest)
  • the manual alignment aka transforms module is well used and important to the user community. The rotation history slider and direct image dragging and rotating are frequent requests. Also addition of scale (& shear)
  • the Fiducial Bspline module is back. After being removed in consideration there are now users that would receive great value added from this. Esp. the category of Registration Repair, i.e. small manual corrections to an (automated) calculated Xform.

Progress: Target Projects

The projects below have been tentatively identified as having registration ties. We will seek discussion with these projects during the week. If your project is listed here, I will try to discuss the registration aspects at some point during the week. Type and length of discussion will vary greatly. Text next to each project below identifies the registration link.

  • Segmentation:
  • Registration
    • Fusion of Anatomy,MRI and Electrophysiology in Parkinson's (Andrzej Przybyszewski):
      • co-registration of subject MRI with anatomical atlas; very small structure target, no intrinsic contrast in MRI
      • involves moving image with drastically clipped FOV, acquisition parameters TBD
      • electrophysiology data avail. as second/validation channel
      • final atlas needs to move into patient-specific stereotactic space, there is no intra-operative scan.
    • Registration with Varying Elastic Parameters (Peter Risholm)
    • COPD Gene Collaboration (Raul San Jose)
      • lung CT datasets for group registration
      • library dataset: in deliberation
  • IGT
    • MRI-guided Robotic Prostate Intervention (Andras Lasso and Junichi Tokuda): Transperineal device: includes registration, where accuracy should be improved; currently the registration is based on a single-slice image; it could possible to improve the accuracy by increasing the number of slices used for the registration.
  • Radiotherapy
  • Analysis
  • Informatics
    • XNAT : discussion with XNAT group re. feasibility of using XNAT central as repository for use-case library.
  • Diffusion
  • NA-MIC Kit Internals

Progress: Notes

  • consider a priority matrix for each case that lists the requirements for robustness, precision, speed. E.g. change tracking applications rank high in precision but low in the other two, IGT applications rank high in speed and robustness etc.
  • MIDAS: discussed with Steve & Casey the use of MIDAS server to host the case library; this would enable additional use of library cases for automated testing (CTest)
  • the manual alignment aka transforms module is well used and important to the user community. The rotation history slider and direct image dragging and rotating are frequent requests. Also addition of scale (& shear)
  • the Fiducial Bspline module is back. After being removed in consideration there are now users that would receive great value added from this. Esp. the category of Registration Repair, i.e. small manual corrections to an (automated) calculated Xform.
  • XNAT Desktop and Slicer Browser: candidate for hosting Case Library

References