Difference between revisions of "2012 Summer Project Week:Overlapping structures"
From NAMIC Wiki
| (4 intermediate revisions by the same user not shown) | |||
| Line 21: | Line 21: | ||
<h3>Approach, Plan</h3> | <h3>Approach, Plan</h3> | ||
| − | + | An initial approach, used in plastimatch, is to represent structure sets as a vector of unsigned characters, with one bit per structure. | |
</div><div style="width: 60%; float: left;"> | </div><div style="width: 60%; float: left;"> | ||
| Line 32: | Line 32: | ||
** Keep original contours as slicer models (view in slices as slice intersections) | ** Keep original contours as slicer models (view in slices as slice intersections) | ||
** Provide routines to rasterize contours into label volumes on demand for editing on a per-structure basis | ** Provide routines to rasterize contours into label volumes on demand for editing on a per-structure basis | ||
| + | ** Caching if needed | ||
** Create vector image representation for running warp operations (warp all structures with single evaluation of pixel deformation function) | ** Create vector image representation for running warp operations (warp all structures with single evaluation of pixel deformation function) | ||
| + | * Caveats | ||
| + | ** Tilted planes are sometimes used (drawing on native MRI planes, but structures defined on CT) | ||
| + | ** Drawing done on saggital or coronal, would be good to see "caps" of structures | ||
| + | ** How to group structure sets (e.g. to exchange with module, CLI)? | ||
</div></div> | </div></div> | ||
| Line 44: | Line 49: | ||
==References== | ==References== | ||
*Greg's progress notes. | *Greg's progress notes. | ||
| − | + | *Mockup of UI (http://www.na-mic.org/Wiki/index.php/2012_Winter_Project_Week:RTSS) | |
</div> | </div> | ||
Latest revision as of 13:55, 22 June 2012
Home < 2012 Summer Project Week:Overlapping structures
Key Investigators
- MGH: Greg Sharp
- Isomics: Steve Pieper
- GE: Jim Miller
- Queen's: Andras Lasso
Objective
We are developing capabilities for Slicer to display and manipulate overlapping structures.
Approach, Plan
An initial approach, used in plastimatch, is to represent structure sets as a vector of unsigned characters, with one bit per structure.
Progress
- Discussions of target data types and capabilities:
- Typically around 100meg for a Head/Neck CT
- Typically no more than 100 structures
- Goal is to provide lossless representation of original data, with conversion from contours to images for editing and warping
- Keep original contours as slicer models (view in slices as slice intersections)
- Provide routines to rasterize contours into label volumes on demand for editing on a per-structure basis
- Caching if needed
- Create vector image representation for running warp operations (warp all structures with single evaluation of pixel deformation function)
- Caveats
- Tilted planes are sometimes used (drawing on native MRI planes, but structures defined on CT)
- Drawing done on saggital or coronal, would be good to see "caps" of structures
- How to group structure sets (e.g. to exchange with module, CLI)?
Delivery Mechanism
This work will be delivered to the NA-MIC Kit as a modification of Slicer interals
References
- Greg's progress notes.
- Mockup of UI (http://www.na-mic.org/Wiki/index.php/2012_Winter_Project_Week:RTSS)
