Difference between revisions of "2014 Project Week:RT FormatConversions"

From NAMIC Wiki
Jump to: navigation, search
 
(One intermediate revision by the same user not shown)
Line 29: Line 29:
 
<div style="width: 27%; float: left; padding-right: 3%;">
 
<div style="width: 27%; float: left; padding-right: 3%;">
 
<h3>Progress</h3>
 
<h3>Progress</h3>
* Csaba prepared a python script (based on Slicer RT) for converting contours to label maps.
+
* Creates python script (based on Slicer RT) for converting contours to label maps (thanks Csaba!)
 
* Reporting module provides conversion from label map to DICOM-SEG
 
* Reporting module provides conversion from label map to DICOM-SEG
* Plastimatch had DICOM-RT import export
+
* Plastimatch has DICOM-RT import export
 
* Will test full roundtrip conversions using QIN lung segmentation challenge datasets
 
* Will test full roundtrip conversions using QIN lung segmentation challenge datasets
 
* Started creating a library of test cases with known issues (donuts, overlapping structures etc)
 
* Started creating a library of test cases with known issues (donuts, overlapping structures etc)

Latest revision as of 04:54, 10 January 2014

Home < 2014 Project Week:RT FormatConversions

Key Investigators

Jayashree Kalpathy-Cramer (MGH)

Andras

Csaba

John Evans (MGH)

Greg Sharp (MGH)

Project Description

Objective

  • Evaluate different options for converting from DICOM-seg/NIFTI/nrrd (mask) to DICOM-RT (contours) and back.
  • Compare performance of Slicer RT/plastimatch/others.

Approach, Plan

  • Develop a test set with range of appearance of structures
    • overlapping
    • donuts

Progress

  • Creates python script (based on Slicer RT) for converting contours to label maps (thanks Csaba!)
  • Reporting module provides conversion from label map to DICOM-SEG
  • Plastimatch has DICOM-RT import export
  • Will test full roundtrip conversions using QIN lung segmentation challenge datasets
  • Started creating a library of test cases with known issues (donuts, overlapping structures etc)
  • Will also create synthetic datasets