Difference between revisions of "Project Week 25/CNN for PseudoCT Generation from T1T2 MR"

From NAMIC Wiki
Jump to: navigation, search
 
(4 intermediate revisions by 3 users not shown)
Line 9: Line 9:
 
*[http://www.imagenglab.com/newsite/salvatore_scaramuzzino/ Salvatore Scaramuzzino] (Magna Graecia University/ASL Vercelli, Italy)
 
*[http://www.imagenglab.com/newsite/salvatore_scaramuzzino/ Salvatore Scaramuzzino] (Magna Graecia University/ASL Vercelli, Italy)
 
*[http://www.imagenglab.com/newsite/mf_spadea/ Maria Francesca Spadea] (Magna Graecia University, Italy)
 
*[http://www.imagenglab.com/newsite/mf_spadea/ Maria Francesca Spadea] (Magna Graecia University, Italy)
*[http://isgwww.cs.uni-magdeburg.de/cas/team.php Gino Gulamhussene] (Universität Magdeburg, Germany)
+
*[http://isgwww.cs.uni-magdeburg.de/cas/team.php Gino Gulamhussene] (University of Magdeburg, Germany)
*[http://isgwww.cs.uni-magdeburg.de/isg/meyer.html Anneke Meyer] (Universität Magdeburg, Germany)
+
*[http://isgwww.cs.uni-magdeburg.de/isg/meyer.html Anneke Meyer] (University of Magdeburg, Germany)
  
 
==Project Description==
 
==Project Description==
Line 26: Line 26:
 
Analysis of the different output with MAE and Bias metrics
 
Analysis of the different output with MAE and Bias metrics
 
|<!-- Progress and Next steps (fill out at the end of project week), bullet points -->
 
|<!-- Progress and Next steps (fill out at the end of project week), bullet points -->
TBD
+
The most suited CNN to be used for this task consisted of an edited U-Net that does not make use of the final Dense Layer, useful for segmentation but not for regression purposes.
 +
Also unpooling layer was used in the decoding part, with zero-filling instead of repetitions for the missing values.
 +
Next planned steps will be the fine tuning of the CNN in order to lower the MAE and BIAS, the translation to TensorFlow (now the code uses Theano as a backend) and the data parallelization.
 +
Interactions with Annette, Gino and Gabriele were important in order to understand how to improve the CNN topology.
 
|}
 
|}
  
 
==Illustrations==
 
==Illustrations==
http://www.slicer.org/img/Slicer4Announcement-HiRes.png
+
[[File:MRI2CT.png]]
  
 
==Background and References==
 
==Background and References==
 
<!-- Use this space for information that may help people better understand your project, like links to papers, source code, or data -->
 
<!-- Use this space for information that may help people better understand your project, like links to papers, source code, or data -->

Latest revision as of 12:11, 30 June 2017

Home < Project Week 25 < CNN for PseudoCT Generation from T1T2 MR


Back to Projects List

Key Investigators

Project Description

This tool allows the user to generate an HU map (Pseudo-CT) from T1/T2 input MRI. The tool is still in the early development stages with good early results in terms of Mean Absolute Error and Bias.

Objective Approach and Plan Progress and Next Steps

The main object is to understand which topology of CNN is the most suited for the Pseudo-CT generation task.

Test of different structures with training on low-res images in order to speed-up computational time. Analysis of the different output with MAE and Bias metrics

The most suited CNN to be used for this task consisted of an edited U-Net that does not make use of the final Dense Layer, useful for segmentation but not for regression purposes. Also unpooling layer was used in the decoding part, with zero-filling instead of repetitions for the missing values. Next planned steps will be the fine tuning of the CNN in order to lower the MAE and BIAS, the translation to TensorFlow (now the code uses Theano as a backend) and the data parallelization. Interactions with Annette, Gino and Gabriele were important in order to understand how to improve the CNN topology.

Illustrations

MRI2CT.png

Background and References