View source for 2014 Summer Project Week: CLI Derived DICOM Data

__NOTOC__
<gallery>
Image:PW-MIT2014.png|[[2014_Summer_Project_Week#Projects|Projects List]]
</gallery>

==Key Investigators==

* Steve Pieper, Isomics
* Andrey Fedorov, BWH
* Jim Miller, GE
* David Clunie, PixelMed (available to answer questions)
* Michael Onken, Open Connections (providing input remotely)

==Project Description==

For QIICR we want to represent all data in DICOM for three specific clinical trials workflows (prostate MR DCE, GBM MR DCT, H&N PETCT).  To make Slicer compatible with this goal, we wish to have a framework to create correct DICOM files for the results of CLI and interactive processing.

<div style="margin: 20px;">
<div style="width: 27%; float: left; padding-right: 3%;">
<h3>Objective</h3>
* See if it's possible to create a wrapper that handles the DICOM I/O for non-DICOM-aware CLIs
* See what extra information from CLIs would be required to know how to wrap them
* See if it makes more sense to have CLIs work more natively with DICOM rather than wrapping them

</div>
<div style="width: 27%; float: left; padding-right: 3%;">
<h3>Approach, Plan</h3>
* Review some sample derived images generated by other code or commercial vendors (David to validate they are correct)
* Develop a reference implementation of DCMTK C++ code that creates derived objects for a simple workflow like denoising of an MR storage object. (Michael to provide code, David to validate correctness)
* Review specific CLIs in Slicer to see what information they need as input and how the output should be represented (Steve and Andrey)
** DCE processing
** image arithmetic including resampling (i.e. images of different orientations)
* Look at how Slicer Execution Model might express the information needed to work with this data (Jim, Andrey, Steve)
 
</div>
<div style="width: 27%; float: left; padding-right: 3%;">
<h3>Progress</h3>
* Held DICOM breakout session and discussed use cases for derived data
* Discussed options for holding data in Subject Hierarchy vs node attributes (prefer Subject Hierarchy wherever possible)
* Michael will deliver DCMTK sample code by the end of June
</div>
</div>