Add doc

docs/index.rst

@@ -43,6 +43,8 @@ See :ref:`Commands <command>` for a description on the avalaible command paramet
 
 If ``-deriv``` is provided, see :ref:`Derivatives <derivatives>` for a descrition of the outputs
 
+Also, check :ref:`Tips and Tricks <tips_tricks>` on how to format images to improve quality of registration between images and segmentation
+
 
 Table of contents
 ******************
@@ -57,6 +59,7 @@ Table of contents
     derivatives
     params
     indiv_params
+    tips_tricks
 
 
 

docs/tips_tricks.rst

@@ -0,0 +1,46 @@
+:orphan:
+
+.. _tips_tricks:
+
+Tips and tricks
+_______________
+
+CT images registration to MRI (be it T1w or T2w) can be somehow tricky.
+
+Here a few tricks:
+
+Reorientation can help
+######################
+
+The -orient LPS corresponds to the direction where the increasing numbers are pointing at (typically RAS in radiological convention) and take the opposite directions (special AFNI...).
+
+If you want to exchange to axis, change "target" letters accordingly
+RPI -> RIP
+
+If you want to revert one dimension, be careful to modify another axis:
+RPI -> RPS becomes LPS to exchange I by S, and good R-L dimension as well
+
+.. code:: bash
+
+    $ 3drefit -orient LPS sub-Phoenix_ses-01_acq-CT_T2star.nii.gz
+
+Possibly followed by
+
+.. code:: bash
+
+    $ fslreorient2std sub-Phoenix_ses-01_acq-CT_T2star.nii.gz sub-Phoenix_ses-01_acq-CT_T2star_reorient.nii.gz
+
+
+Cropping (both T1w et T2w) can help
+###################################
+
+Include the full head for the CT image
+
+Be careful with Left-Right orientation
+######################################
+
+Most of the images in our center are acquired in the right direction, but with an encoding matching inverse Left-Right Encoding to the one that is coming out of the MRI
+
+.. code:: bash
+
+    $ 3drefit -orient LPI sub-Rusty/ses-01/anat/sub-Rusty_ses-01_acq-CT_run-02_T2star.nii.gz