Hi Marcella, thank you for your report!
>I've done with DICOM Import...I don't know if it is an error: the bottom bar >stops at 66% of the loading. I've tried with 2 different computers. However >it works and it estimates the diffusion tensor. great! Thanks for letting me know. There is a bug for the percentage issue: http://bugs.mitk.org/show_bug.cgi?id=9967 > -ITK Tensor Reconstruction: it seems working > -Estimate Diffusion Image from Tensors: it works with # gradient directions =12 (I have more but less than 42) Once you have estimated tensors you can go back and estimate any number of gradient directions from the tensors. This was implemented since the fiber tracking requires qballs, and qball-reconstruction works more stable with a higher number of gradient directions distributed on the sphere. >- Estimate Q-Ball Image from Tensors ...I had an error, probably for memory >problems Yes, the direct way is memory intensive. Going from tensors to diffusion image to qballs should do the job. >- ODF details and ODF glyphs from the reconstructed tensor: ok >- Reconstruction of Q-Balls with all the methods: it doesn't produce any >Data, maybe doesn't it finish for the same memory problems (I do not have >errors)? What's your setup? windows? 32bit? >Could I ask you what is "Correct DKFZ Gradient Scheme"? This is a special flag to correct corrupt images that were acquired many years ago, with wrong b-matrix information in the dicom files. Actually, this flag should be removed since it is not of general use. >I've not tried with FSL again (I do not have the bvecs and bval files, how >are they? Like B-matrix files?). In my previous email I provided some details about this for a specific example of 6 directions. Here is a copy: 3 files: diffusionimage.fsl, diffusionimage.fsl.bvals, and diffusionimage.fsl.bvecs with the bvals-file containing 0 1000 1000 1000 1000 1000 1000 and the bvecs-file containing 0 0.7103 -0.7103 0.002857 0.002856 0.708873 -0.708873 0 0.000708 0.000708 -0.708531 -0.708174 -0.705331 -0.705331 0 -0.703899 -0.703899 -0.705674 0.706032 -0.002833 -0.002833 >Thanks >Marcella Best regards Klaus -----Messaggio originale----- Da: Fritzsche, Klaus H. [mailto:[email protected]] Inviato: giovedì 17 novembre 2011 17.36 A: Marcella Laganà Cc: [email protected] Oggetto: AW: about MITK Diffusion Hi Marcella, For diffusion images, you should use the "DICOM Import" view, not the "File Open" or drag n drop. The open dialog will simply trigger the ITK-mechanism for loading files, which will not result in appropriate extraction of the b-matrix information. Sorry about this confusion. You mentioned, that you have 4D fsl-files, probably <your-filename>.nii You could also rename the file to <your-filename>.fsl and create an accompanying <your-filename>.fsl.bvecs and a <your-filename>.fsl.bvals file to load the dataset. Please let me know, if you succeeded. Best Klaus ________________________________________ Von: Marcella Laganà [[email protected]] Gesendet: Donnerstag, 17. November 2011 15:09 An: Fritzsche, Klaus H. Betreff: R: about MITK Diffusion Thank you. The screenshot I've sent you has been done with the first method, so loading the dicom. I've tried again with another patients of a different protocol (12 directions and 1 b0, 2 averages), but ITK tensor reconstruction doesn't become active. I have 1 object in the Data Manager panel and changing the "Time" point I see the different directions. First, I wanted to average the 2 runs with Preprocessing--> Reduce Size --> average redundant gradients, but also that command is not available. Thanks again Marcella -----Messaggio originale----- Da: Fritzsche, Klaus H. [mailto:[email protected]] Inviato: martedì 15 novembre 2011 23.17 A: Marcella Laganà Cc: [email protected] Oggetto: AW: about MITK Diffusion Hi Marcella, you have three different options to load diffusion weighted datasets. 1) dicom import of Siemens files If you choose one of the following two options, you have to change the file-endings. If you don't do that, MITK will not be able to differentiate them form plain 4D-images. After loading, you will recognize the different icon used for diffusion images, compared to plain 4D images: 2) use the nrrd-format, like Slicer does first rename *.nhdr => *.hdwi *.nrrd => *.dwi 3) use the nifti-format, like FSL does first rename *.nii => *.fsl *.nii.gz => *.fslgz For option 2), the nrrd-header should include DWMRI_b-value, DWMRI_gradient_xxxx, measurement frame, modality:=DWMRI, as in the following example: NRRD0004 type: short dimension: 4 space: left-posterior-superior sizes: 7 32 66 20 space directions: none (1,0,0) (0,1,0) (0,0,1) kinds: vector domain domain domain endian: little encoding: raw space origin: (0,0,0) data file: data.raw DWMRI_b-value:=1000.000000 DWMRI_gradient_0000:=0.000000 0.000000 0.000000 DWMRI_gradient_0001:=0.710300 0.000708 -0.703899 DWMRI_gradient_0002:=-0.710300 0.000708 -0.703899 DWMRI_gradient_0003:=0.002857 -0.708531 -0.705674 DWMRI_gradient_0004:=0.002856 -0.708174 0.706032 DWMRI_gradient_0005:=0.708873 -0.705331 -0.002833 DWMRI_gradient_0006:=-0.708873 -0.705331 -0.002833 measurement frame:=(1.000000,0.000000,0.000000) (0.000000,1.000000,0.000000) (0.000000,0.000000,1.000000) modality:=DWMRI For option 3), the *.fsl (or *.fslgz) file must be accompanied with a bvals and a bvecs file for the b-matrix information: diffusionimage.fsl diffusionimage.fsl.bvals diffusionimage.fsl.bvecs with the bvals-file containing 0 1000 1000 1000 1000 1000 1000 and the bvecs-file containing 0 0.7103 -0.7103 0.002857 0.002856 0.708873 -0.708873 0 0.000708 0.000708 -0.708531 -0.708174 -0.705331 -0.705331 0 -0.703899 -0.703899 -0.705674 0.706032 -0.002833 -0.002833 using the same example as in 2). Hope, this helps! Regards Klaus ________________________________________ Von: Marcella Laganà [[email protected]] Gesendet: Dienstag, 15. November 2011 22:17 An: Fritzsche, Klaus H. Betreff: R: about MITK Diffusion Thank you for your fast answer. This software seems very comprehensive, I've read the documentation but I'm not able to start using it. I've tried again loading dicom, or different 3D volumes, or 1 4D volume (fsl output). I've also tried creating a 4D volume which is composed by only 1 b0 and 12 directions, so that I've selected the option of 12 directions. I've clicked on Tensors and then on the dataset, but the ITK tensor reconstruction didn't become active (attached the screenshot). Is it the same to load separate 3D files or 4D or dicom? Don't I have to insert the B-matrix? Thanks and regards Maria Marcella Laganà ------------------------------------------------------------------------------ Cloud Services Checklist: Pricing and Packaging Optimization This white paper is intended to serve as a reference, checklist and point of discussion for anyone considering optimizing the pricing and packaging model of a cloud services business. Read Now! http://www.accelacomm.com/jaw/sfnl/114/51491232/ _______________________________________________ mitk-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mitk-users
