Difference between revisions of "Slicer3:Module:Rician Noise Removal"

From NAMIC Wiki
Jump to: navigation, search
m (Text replacement - "http://www.slicer.org/slicerWiki/index.php/" to "https://www.slicer.org/wiki/")
 
(6 intermediate revisions by 3 users not shown)
Line 1: Line 1:
===Module Name===
+
<big>'''Note:''' We are migrating this content to the slicer.org domain - <font color="orange">The newer page is [https://www.slicer.org/wiki/Slicer3:Module:Rician_Noise_Removal  here]</font></big>
Rician Noise Removal in Diffusion Tensor MRI (DWI and tensors)
 
 
 
{|
 
|[[Image:screenshotBlank.png|thumb|280px|Caption 1]]
 
|[[Image:screenshotBlank.png|thumb|280px|Caption 2]]
 
|[[Image:screenshotBlank.png|thumb|280px|Caption 3]]
 
|}
 
 
 
== General Information ==
 
===Module Type & Category===
 
 
 
Type: CLI
 
 
 
Category: Filtering DWI and tensors
 
 
 
===Authors, Collaborators & Contact===
 
* Saurav Basu: University of Utah
 
* Thomas Fletcher, University of Utah
 
* Ross Withaker, University of Utah
 
* Contact: Thomas Fletcher
 
 
 
===Module Description===
 
Rician noise introduces a bias into MRI measurements that
 
can have a significant impact on the shapes and orientations of ten-
 
sors in diffusion tensor magnetic resonance images. This is less of a
 
problem in structural MRI, because this bias is signal dependent and
 
it does not seriously impair tissue identification or clinical diagnoses.
 
However, diffusion imaging is used extensively for quantitative evalua-
 
tions, and the tensors used in those evaluations are biased in ways that
 
depend on orientation and signal levels. This paper presents a strat-
 
egy for filtering diffusion tensor magnetic resonance images that ad-
 
dresses these issues. The method is a maximum a posteriori estima-
 
tion technique that operates directly on the diffusion weighted images
 
and accounts for the biases introduced by Rician noise. We account for
 
Rician noise through a data likelihood term that is combined with a
 
spatial smoothing prior. The method compares favorably with several
 
other approaches from the literature, including methods that filter dif-
 
fusion weighted imagery and those that operate directly on the diffusion
 
tensors.
 
 
 
== Usage ==
 
 
 
===DWI filtering===
 
 
 
====Examples, Use Cases & Tutorials====
 
 
 
USAGE
 
--------------
 
dwiFilter <arguments>
 
Arguments:
 
1. Input File Name
 
2. Output File Name
 
3. NumIterations
 
4. Conductance
 
5. TimeStep
 
6. Filter Type : (Simple Aniso-0,Chi Squared-1,Rician-2,Gaussian-3)
 
7. Sigma for bias correction
 
8. Lamda (Rician Correction Term)
 
9. Lamda (Gaussian Correction Term)
 
 
 
Argument Description:
 
 
 
<Input File Name>
 
Name of the DWI file to be filtered. For example
 
<noisyDWI_10.nhdr> is a noisy DWI file provided
 
in the data directory. It was generated by adding
 
synthetic Rician noise with a sigma=10 to a cleanDWI.nhdr
 
 
 
<Output File Name>
 
Name of the filtered DWI file. For example
 
<filteredDWI.nhdr>
 
 
 
<NumIterations>
 
Number of iterations you want to run the filter for.
 
 
 
<Conductance>
 
The value of the conductance term in anisotropic
 
diffusion filtering (Ex: 1.0)
 
Note: Large Conductance will oversmooth the image
 
It is important to tune the conductance to obtain
 
best results.
 
 
 
<Time Step>
 
This determines the step size in the gradient
 
descent. It can be atmost 0.0625.
 
 
 
<Filter Type>
 
Can Take 3 values:
 
0 means perform simple anisotropic diffusion
 
 
 
* - 1 means perform Chi-Squared smoothing (square the image and perform anisotropic diffusion and then subtract the variance of the noise, and take square root. (The square of a Rice distribution is a Chi Squared distribution with known bias equal to the variance of the noise) (Refer:Max likelihood Est. of Rician Ditribution Parameters. Sijbers et. al)
 
* - 2 means Perform Rician bias correction filtering.(Refer: Rician Noise Removal in DT-MRI.)
 
* - 3 is same as 2 except use a Gaussian Attachment Term .
 
 
 
<Sigma>
 
Estimate of noise in the data.
 
This can be done by squaring the airvoxels
 
in the real data. The sum of square of all
 
the intensities in the air region should equal
 
2*variance of the noise in the data.
 
(Sijbers et. al)
 
 
 
<lamda1, lamda2>
 
The weights for the Rician and Gaussian
 
attachment terms.
 
 
 
EXAMPLE
 
-------------
 
dwiFilter ../data/noisyDWI_10.nhdr filteredDWI.nhdr 1 1.0 0.0625 2 10 100 0
 
 
 
Filters the noisyDWI_10.nhdr for 1 iteration with a conductance of 1.0
 
timeStep 0.0625 using Rician filtering with a Rician attachement term
 
weight of 100. The estimate of noise in the input image is a sigma of 10
 
The filtered image is filteredDWI.nhdr.
 
 
 
===Tensor filtering===
 
 
 
Usage
 
--------------
 
tensorDiffuse <Arguments>
 
1. FilterType:(0-Euclidean, 1-Log Space,2-Riemannian)
 
2. numIterations:Iterations For Anisotropic Diffusion
 
3. timeStep:timeStep Used in Anisotropic Diffusion
 
4. conductance:Conductance used for Anisotropic Diffusion
 
5. Input (filename of input data)
 
6. Output (filename of output data)
 
 
 
Arguments 2,3,4 have the same meaning as described for dwiFilter (see above).
 
 
 
Argument 1 describes the filter type
 
* - 0: Euclidean Space filtering (tensors are treated as 6-d vectors)
 
* - 1: Log Space filtering (Fast and Simple Calculus on Tensors in the Log-Euclidean Framework. In J. Duncan and G. Gerig, editors, Proceedings of the 8th Int. Conf. on Medical Image Computing and Computer-Assisted Intervention - MICCAI 2005, Part I, volume 3749 of LNCS, Palm Springs, CA, USA, October 26-29, pages 115-122, 2005. Springer Verlag)
 
* - 2. Riemannian Space Filtering(A Riemannian Framework for the Processing of Tensor-Valued Images. In Ole Fogh Olsen, Luc Florak, and Arjan Kuijper, editors, Deep Structure, Singularities, and Computer Vision (DSSCV), number 3753 of LNCS, pages 112-123, June 2005. Springer Verlag.)
 
 
 
Currently, the Riemannian filter adjustment for negative eigen-values
 
is hard-coded in the source file.
 
 
 
Argument 5 is the name of the noisyTensor input.
 
Argument 6 is the name of the output tensor file
 
 
 
EXAMPLE
 
--------------
 
tensorFilter 2 1 0.0625 1.0 noisyTensor_10.nhdr FilteredTensor.nhrd
 
 
 
 
 
===Quick Tour of Features and Use===
 
List all the panels in your interface, their features, what they mean, and how to use them. For instance:
 
 
 
* '''Input panel:'''
 
* '''Parameters panel:'''
 
* '''Output panel:'''
 
* '''Viewing panel:'''
 
 
 
== Development ==
 
 
 
===Known bugs===
 
 
 
Follow this link to the Slicer3 bug tracker:
 
 
 
http://na-mic.org/Mantis/main_page.php
 
 
 
===Usability issues===
 
 
 
Follow this link to the Slicer3 bug tracker. Please select the '''usability issue category''' when browsing or contributing:
 
 
 
http://na-mic.org/Mantis/main_page.php
 
 
 
===Source code & documentation===
 
 
 
Customize following links for your module:
 
 
 
http://www.na-mic.org/ViewVC/index.cgi/
 
 
 
Links to documentation generated by doxygen:
 
 
 
http://www.na-mic.org/Slicer/Documentation/Slicer3/html/
 
 
 
== More Information ==
 
 
 
===Acknowledgement===
 
This work is part of the National Alliance for Medical Image Computing
 
(NAMIC), funded by the National Institutes of Health through the NIH Roadmap
 
for Medical Research, Grant U54 EB005149. Information on the National Centers
 
for Biomedical Computing can be obtained from http://nihroadmap.nih.gov/
 
bioinformatics. Funding for this work has also been provided by Center for
 
Integrative Biomedical Computing, NIH NCRR Project 2-P41-RR12553-07. We
 
thank Weili Lin and Guido Gerig from the University of North Carolina for
 
providing us with the DW-MRI data. Glyph visualizations created with Teem
 
(http://teem.sf.net).
 
 
 
 
 
===References===
 
Basu, S., Fletcher, P.T., Whitaker, R.T. Rician Noise Removal in Diffusion Tensor MRI. In Medical Image Computing and Computer Assisted Intervention (MICCAI), LNCS 4190, pp. 117-125, October, 2006.
 
[[http://www.sci.utah.edu/~fletcher/BasuDTIFilteringMICCAI2006.pdf | Paper link ]]
 

Latest revision as of 17:26, 10 July 2017

Home < Slicer3:Module:Rician Noise Removal

Note: We are migrating this content to the slicer.org domain - The newer page is here