From NITRC Wiki

RESTORE: Robust estimation of tensors by outlier rejection

Robust estimation of tensors by outlier rejection (RESTORE) is a popular method of tensor estimation. Here we show how the implementation from CAMINO can be used directly within JIST. http://www.cs.ucl.ac.uk/research/medic/camino/

Prerequisites

• JIST Pre-Beta 1.3 or Newer: http://www.nitrc.org/frs/?group_id=228
• NOTE: For Pre-Beta 1.3 and below, use the .layout files from August 2009. For Beta 1.4 and above, use the layout files from October 2009.

References

• Chang L-C, Jones DK and Pierpaoli C, RESTORE: Robust estimation of tensors by outlier rejection, Magnetic Resonance in Medicine, 53(5), 1088-1095, 2005. http://www3.interscience.wiley.com/cgi-bin/fulltext/110474029/PDFSTART
• P. A. Cook, Y. Bai, S. Nedjati-Gilani, K. K. Seunarine, M. G. Hall, G. J. Parker, D. C. Alexander, "Camino: Open-Source Diffusion-MRI Reconstruction and Processing", 14th Scientific Meeting of the International Society for Magnetic Resonance in Medicine, Seattle, WA, USA, p. 2759, May 2006. http://www3.interscience.wiley.com/journal/120173564/abstract http://www.cs.ucl.ac.uk/research/medic/camino/index.htm

Example

Method

• Download sample data here: http://www.nitrc.org/frs/download.php/1100/RESTORE-test.zip
• Download the layout file for RESTORE here : http://www.nitrc.org/frs/?group_id=228
• Load the layout file. Note: This layout file compares RESTORE tensor estimation to LINEAR tensor estimation.

• Select the "DTI Data" Source
  • Select the "DTIFiles.txt" file that was included with the sample data.
  • Edit the DTIFiles.txt in a text editor and change the paths to point to the downloaded images.
• Select the "3D Mask Data" Source
  • Select the "MaskFiles.txt" file that was included with the sample data.
  • Edit the MaskFiles.txt in a text editor and change the paths to point to the downloaded images.
  • Note that is it very important to specify a conservative mask (not too many background pixels) for RESTORE because the method is VERY slow when the tensor is poorly defined (high noise).
• Select the "Noise Level" Source
  • Select the "NoiseLeve.txt" file that was included with the sample data.
  • Here we have set the estimate noise level (from jist:Noise_Field_Estimation ) to 14.
  • You may adjust this parameter to change the estimation properties.
• Select "Define Scheme Modules"
  • Set the "table of diffusion weighting directions" to the "grads.dpf" file that was included with the sample data.
    • This file contains a list of diffusion weighting directions for each 3D volume within the two 4D volumes (Image1 and Image2)
  • Set the "table of b-values" to the "bs.b" file that was included with the sample data.
    • This file contains a list of the b-vale for each 3D volume within the two 4D volumes (Image1 and Image2).
  • Set the "big delta" to the "bigDelta.txt" file that was included with the sample data.
    • This file contains a dummy value for the diffusion time as it is not used in RESTORE.
  • Set the "small delta" to the "smallDelta.txt" file that was included with the sample data.
    • This file contains a dummy value for the length of the diffusion pulse as it is not used in RESTORE.
  • Set the "te" to the "te.txt" file that was included with the sample data.
    • This file contains a dummy value for the echo time as it is not used in RESTORE.
• Select Project->Layout preferences and specify an output directory.
• Select Project->Global preferences.
  • Ensure that you have selected the MIPAV JRE's for the java binary.
  • Otherwise, you must ensure that the MIPAV numerical libraries are on your default java classpath.
• Save the layout.
• Run the layout in the process manager.