We will go through the steps for prcoessing the MRI data and getting it ready for making measurements in mimics.
Note: You can find all the files in this GitHub repository itself. Note that the files here have hard-coded directories paths and subject lists. You can change those as per your requirements.
- You should download ITK-SNAP from here.
- You also need the command line tool
c3dwhich comes alomg with ITK-SNAP and can learn more about it here. - Matlab
- I ran all the code on a mac.
- I had received the data from Dillan and Zach in the form of zip file as
SAyr3BF001.zip. I take those kind of files and then apply the following steps:
Unzip the downloaded zip file in the form, for example,SAyr3BF001.zip. See the file unzip_all_subjects.sh. Download a sample zip file from here to work through this tutorial.
Now you will see a lot of folders within the extracted folder. And most of these folders have a lot of dicom files. We will try tp find the correct folder which hhas the T1w image. Use this convert_to_nifti.sh script which will find the folder which has the T1w image and will then convert the dicom series into a usuable nifti file.
Once, you get one nifti file for each subject, then you are ready to create the template image in nifti format. Note that is a separate process which I will discuss in the Template building. Refer to that section and then continue to Step 4. For now, let's say that the created template image is named as template.nii.gz. Normalize the intensity range of the image template.nii.gz using skull_norm.sh.
Now, we can threshold the template image using c3d to obtain a binary skull iamge, let's name it as template_binary.nii.gz. Refer to the documentation for more here and described here:
Syntax: -thresh <u1 u2 vIn vOut>
Thresholds the image, setting voxels whose intensity is in the range [u1,u2] to vIn and all other voxels to vOut. Values u1 and u2 are intensity specifications (see below). This means that you can supply values inf and -inf for u1 and u2 to construct a one-sided threshold. You can also specify u1 and u2 as percentiles.
c3d in.img -threshold -inf 128 1 0 -o out.img
c3d in.img -threshold 64 128 1 0 -o out.img
c3d in.img -threshold 20% 40% 1 0 -o out.img
Choose the -threshold as per your judgement by opening up the template.nii.gz image and then see the values above which you want to set it to foreground and the ones which you want to set as background.
Lastly, we need to convert the template_binary.nii.gz to a dicom series using the file dicomwritevolume.m. We need to convert it to a dicom series because that is the only format which Mimics likes.
First, read the image using niftiread. See documentation here.
V = niftiread('template_binary.nii.gz');
The, use the dicomwritevolume.m and run it as:
fname = '/path/to/some/folder/subject.zip'
dicomwritevolume(fname, V, VS)You will then have the /path/to/some/folder/subject.zip file which has the dicom series. Now, you open that file in mimics and then do the measurements!
Here, we will build the template using a bunch of T1w images in nifti files. First, remember to normalize each of the images, that are being used to build the template, using the commands in skull_norm.sh.
- You also need to have
greedyinstalled. See instructions and documentation here. - You also need some bianaries in order to build the template. Get the binaries from this box link.
- You need the files in the folder
template_building_filesin the GitHub repository:manifest.csv: here you provide the complete path to each of the inout image, and also assign a name to each image.flags.json: here you can provide different options for the template to be built. For example. here we care about either if the template building proccess is intialized using the average of all the input image or whether a particular image actts as an intial template image.greedy_build_template.shis the main template building script using the greedy tool. Go through it to understand the algorithm but you should not edit this file. Lastly,pybatch.shis a file so that you can run on clusters; again you do not need to edit this file.
And then run the following command:
export PATH="/path/to/where/you/have/the/greedy/binaries/":$PATH
sh greedy_build_template.sh -p flags.json -i manifest.csv -o template_work
template_work is where all the logs and the final template will be stored.
References: Read the following paper to understandhow templates are generally built.
Unbiased diffeomorphic atlas construction for computational anatomy by Sarang Joshi, Brad Davis, Matthieu Jomier, Guido Gerig.