TubeTK: Difference between revisions

From KitwarePublic
Jump to navigationJump to search
No edit summary
 
(26 intermediate revisions by 5 users not shown)
Line 3: Line 3:


{| border="1" cellpadding="10" cellspacing="0"
{| border="1" cellpadding="10" cellspacing="0"
|+ [[image:TubeTK_Header.jpg|1000px]]
|+ [[image:TubeTK_Header.jpg|1000px|TubeTK]]
|-
|-
| style="background:#efefef;" align="left" valign="top" width="150px" |  
| style="background:#efefef;" align="left" valign="top" width="150px" |  
'''[[TubeTK|Home]]'''
*[[TubeTK/About|About]]
*[[TubeTK/Images|Image Gallery]]
*[[TubeTK/Data|Data and Publications]]
<br>
----
<br>
'''For Users'''
* [[TubeTK/Installation|Installation]]
* [[TubeTK/Documentation|Methods & Apps]]
* [[TubeTK/Slicer|TubeTK with 3D Slicer]]
* [[TubeTK/OsiriX|TubeTK with OsiriX]]
<br>
<br>
<b>[http://public.kitware.com/Wiki/TubeTK Home]</b>
----
*[[TubeTK/About | About]]
*[[TubeTK/Events | Events]]
*[[TubeTK/Images | Image Gallery]]
*[[TubeTK/Data | Data and Publications]]
<br>
<br>
<hr>
'''For Developers'''
* [[TubeTK/Development|Development Docs]]
<br>
<br>
<b>For Users</b>
----
* [[TubeTK/Documentation | Methods & Apps]]
* [http://tubetk.982995.n3.nabble.com Support Forum]
* [[TubeTK/Slicer | TubeTK with Slicer]]
* [[TubeTK/ImageJ | TubeTK with ImageJ]]
<br>
<br>
<hr>
'''[https://github.com/TubeTK/TubeTK/issues Report Bugs<br>Request Features]'''
<br>
<br>
<b>For Developers</b>
* [[TubeTK/Development | Development Docs]]
<br>
<br>
<hr>
----
<br>
<br>
<b>[[TubeTK/Team | Contact Us]]</b>
'''[[TubeTK/Contact|Contact Us]]'''


| align="left" width="800px" |  
| align="left" width="800px" |  
= Overview =
= Overview =


TubeTK is being developed to host algorithms for applications involving images of tubes.   By focusing on the geometry of tubes we can accomplish grand challenges in image analysis in a subset of significant cases.
TubeTK is being developed to host algorithms for applications involving images of tubes (blood vessel in medical images, roads in satellite images, etc.).  It also offers methods for handling other geometries (points, surfaces, and densities) in images.  
 
By focusing on local geometric structure, the algorithms are able to accomplish segmentations, registrations, and other analyses that consider the physicial properties of objects and their variations, while not requiring limiting assumptions on the specific arrangement or general shape of the objects in the images.  We are applying these techniques to push image understanding in new directions such as:
# registration of abdominal images even when organs slides against one another
# forming statistical atlases of intra-canrial vessel network topology even when that topology changes between subjects
# segmentation of arbitrary objects in images even when intensity statistics of those objects, and the objects around them, vary from image to image.
 
At this time TubeTK is targeted for
# Software developers who wish to write code to integrate our algorithms into their applications
# Researchers who can write bash and other scripts to string together TubeTK's command-line tools
 
TubeTK offers various interface layers:
* '''TubeTK/Base:''' This is the algorithms library. It is the lowest level of access to the methods of TubeTK. It is only available via C++, and it requires considerable expertise to effectively combine and call its methods to do anything useful. Iterfacing directly with these algorithms is not recommended and is not well supported. Unit-level testing is performed continuously on these methods.
* '''TubeTK/ITKModules:''' This is the ITK interface to select methods in TubeTK/Base. This level of interface is intended for ITK users and Python scripts writers. The methods exposed represent a level of modularization that invites experimentation, integration with other toolkits (e.g., Scikit-Learn), and development of processing pipelines that accomplish significant image analysis goals. The interface is available as an ITK Extension and thereby available via Python using Wrapped ITK.
* '''TubeTK/Applications:''' These are the command-line interface (CLI) equivalents to the methods available via TubeTK/ITKModules. This is intended for bash, bat, and other system-call scripts. The level of modularization and intended users are similar to those of TubeTK/ITKModules. C++ and python-based CLIs are provided. Continuous, unit-level testing of TubeTK/ITKModules is provided via these applications.
* '''TubeTK/Experiments:''' These are Python Jupyter notebooks that combine many TubeTK/ITKModules into Python scripts that show how to accomplish high-level image analysis goals with TubeTK. They are intended to be an interactive basis for exploring TubeTK. Python and Jypter notebooks packages must be installed on your computer to run these. These can also be (and are) run as tests to check performance (whereas the unit-level tests focus on regression).
* '''TubeTK/SlicerModules:''' These are Slicer modules that combine many of the TubeTK/ITKMoudles into Slicer elements that accomplish select high-level image analysis tasks using TubeTK.
 
If you have questions regarding or suggestions for improving TubeTK, please do not hesitate to [[TubeTK/Contact|contact the development team]].


== Features ==
== Features ==
Line 49: Line 71:
* Vascular atlas formation
* Vascular atlas formation
* Radiofrequency ablation guidance via intra-operative registration of pre-op CT with intra-op Ultrasound
* Radiofrequency ablation guidance via intra-operative registration of pre-op CT with intra-op Ultrasound
* [[TubeTK/Tumor Micro Environments | Quantifying tumor micro-environments (with UNC BME Dept.)]]
* Quantifying tumor micro-environments


== Technical Focus ==
== Technical Focus ==
Line 56: Line 78:
* [[TubeTK/Sliding_Organ_Registration | Sliding organ registration]] are methods for registering images of multiple organs in which the organs may have shifted, expanded, or compressed independently.
* [[TubeTK/Sliding_Organ_Registration | Sliding organ registration]] are methods for registering images of multiple organs in which the organs may have shifted, expanded, or compressed independently.
* [[TubeTK/Intra-operative_Ultrasound_Registration | Intra-operative ultrasound registration]] is the grand challenge of real-time transcription of pre-operative surgical plans into intra-operative ultrasound images.
* [[TubeTK/Intra-operative_Ultrasound_Registration | Intra-operative ultrasound registration]] is the grand challenge of real-time transcription of pre-operative surgical plans into intra-operative ultrasound images.
|-
 
|}
== Acknowledgement ==
 
If you find TubeTK useful for your work and publications, please include a reference to this website and to
* S. R. Aylward and E. Bullitt, “Initialization, noise, singularities, and scale in height ridge traversal for tubular object centerline extraction,” Medical Imaging, IEEE Transactions on, vol. 21, no. 2, pp. 61–75, 2002.
Thank you!
 
== External Links ==
 
* [http://open.cdash.org/index.php?project=TubeTK TubeTK Dashboard]
* [https://github.com/KitwareMedical/TubeTK/issues TubeTK GitHub Issues]
* [https://github.com/KitwareMedical/TubeTK TubeTK GitHub Repository]
* [http://midas3.kitware.com/midas/community/7 TubeTK Midas Platform Community]
* [https://github.com/KitwareMedical KitwareMedical GitHub]
 
 
[[Category:TubeTK|*]]

Latest revision as of 12:31, 9 May 2016

__NOTITLE__


TubeTK

Home




For Users




For Developers




Report Bugs
Request Features




Contact Us

Overview

TubeTK is being developed to host algorithms for applications involving images of tubes (blood vessel in medical images, roads in satellite images, etc.). It also offers methods for handling other geometries (points, surfaces, and densities) in images.

By focusing on local geometric structure, the algorithms are able to accomplish segmentations, registrations, and other analyses that consider the physicial properties of objects and their variations, while not requiring limiting assumptions on the specific arrangement or general shape of the objects in the images. We are applying these techniques to push image understanding in new directions such as:

  1. registration of abdominal images even when organs slides against one another
  2. forming statistical atlases of intra-canrial vessel network topology even when that topology changes between subjects
  3. segmentation of arbitrary objects in images even when intensity statistics of those objects, and the objects around them, vary from image to image.

At this time TubeTK is targeted for

  1. Software developers who wish to write code to integrate our algorithms into their applications
  2. Researchers who can write bash and other scripts to string together TubeTK's command-line tools

TubeTK offers various interface layers:

  • TubeTK/Base: This is the algorithms library. It is the lowest level of access to the methods of TubeTK. It is only available via C++, and it requires considerable expertise to effectively combine and call its methods to do anything useful. Iterfacing directly with these algorithms is not recommended and is not well supported. Unit-level testing is performed continuously on these methods.
  • TubeTK/ITKModules: This is the ITK interface to select methods in TubeTK/Base. This level of interface is intended for ITK users and Python scripts writers. The methods exposed represent a level of modularization that invites experimentation, integration with other toolkits (e.g., Scikit-Learn), and development of processing pipelines that accomplish significant image analysis goals. The interface is available as an ITK Extension and thereby available via Python using Wrapped ITK.
  • TubeTK/Applications: These are the command-line interface (CLI) equivalents to the methods available via TubeTK/ITKModules. This is intended for bash, bat, and other system-call scripts. The level of modularization and intended users are similar to those of TubeTK/ITKModules. C++ and python-based CLIs are provided. Continuous, unit-level testing of TubeTK/ITKModules is provided via these applications.
  • TubeTK/Experiments: These are Python Jupyter notebooks that combine many TubeTK/ITKModules into Python scripts that show how to accomplish high-level image analysis goals with TubeTK. They are intended to be an interactive basis for exploring TubeTK. Python and Jypter notebooks packages must be installed on your computer to run these. These can also be (and are) run as tests to check performance (whereas the unit-level tests focus on regression).
  • TubeTK/SlicerModules: These are Slicer modules that combine many of the TubeTK/ITKMoudles into Slicer elements that accomplish select high-level image analysis tasks using TubeTK.

If you have questions regarding or suggestions for improving TubeTK, please do not hesitate to contact the development team.

Features

  • Centerline vascular segmentation
  • Vascular atlas formation
  • Vascular network to image registration
  • Organ segmentation
  • Vascular-ness measures
  • Multi-modality support

Driving Applications

  • NeuralNav: Brain tumor resection guidance via intra-operative registration
  • Diagnosis of retinopathy of prematurity
  • Analysis of vascular differences within and between populations to identify focal (e.g., stroke) and diffuse (e.g., schizophrenia) diseases.
  • Vascular atlas formation
  • Radiofrequency ablation guidance via intra-operative registration of pre-op CT with intra-op Ultrasound
  • Quantifying tumor micro-environments

Technical Focus

  • Vascular pattern analysis is the characterization and comparison of individuals and populations based on "spatial graphs" as representations of vascular networks.
  • Sliding organ registration are methods for registering images of multiple organs in which the organs may have shifted, expanded, or compressed independently.
  • Intra-operative ultrasound registration is the grand challenge of real-time transcription of pre-operative surgical plans into intra-operative ultrasound images.

Acknowledgement

If you find TubeTK useful for your work and publications, please include a reference to this website and to

  • S. R. Aylward and E. Bullitt, “Initialization, noise, singularities, and scale in height ridge traversal for tubular object centerline extraction,” Medical Imaging, IEEE Transactions on, vol. 21, no. 2, pp. 61–75, 2002.

Thank you!

External Links