Resample images to desired resolution #85
Conversation
config_sct_run_batch.yml
Outdated
| @@ -1,6 +1,6 @@ | |||
| # config file for sct_run_batch | |||
| path_data: /scratch/pabaua/data-multi-subject-p | |||
| path_output: /scratch/pabaua/results_csa_t2_2 | |||
| path_output: /scratch/pabaua/results_csa_t2_3 | |||
There was a problem hiding this comment.
why these changes? they are local (ie specific to your config)
There was a problem hiding this comment.
It is to keep track of the configs i used when i run jobs on compute canada. I will not merge this into master. In the future i will fork the repo for these kinds of PRs.
jcohenadad
changed the title
process_data.sh
Outdated
| if [ -e $FILESEGMANUAL ]; then | ||
| echo "Found! Using manual segmentation." | ||
| rsync -avzh $FILESEGMANUAL ${FILESEG}.nii.gz | ||
| sct_resample -i ${FILESEGMANUAL}.nii.gz -mm $interp -x nn -o ${FILESEGMANUAL}_r.nii.gz | ||
| rsync -avzh ${FILESEGMANUAL}.nii.gz ${FILESEG}.nii.gz |
There was a problem hiding this comment.
problem here, the resampled ${FILESEGMANUAL}_r.nii.gz should be copied to ${FILESEG}.nii.gz
There was a problem hiding this comment.
Suggestion was implemented in commit: 482a704
…into t2w_resample
- update file name resample
process_data.sh
Outdated
| @@ -113,7 +114,7 @@ cd anat | |||
| # Reorient to RPI and resample file | |||
| if [ $contrast == "t2" ]; then | |||
| contrast_str="T2w" | |||
| interp="0.8x0.8x0.8" | |||
| interp="1x1x1" | |||
There was a problem hiding this comment.
i was suggesting to resample to 1mm iso only for "checking" the effect of this resampling, not for the "stable" release of the pipeline
There was a problem hiding this comment.
Sorry for the confusion, i was thinking of re-running this PR on Compute Canada to put results in the article's supplementary material (It could partly explain differences observed between t1 and t2 results). @jcohenadad, what are your thoughts on that?
There was a problem hiding this comment.
sounds good. IN fact, if we explore the effect of resampling, we probably want to also try other resamplings, e.g. 1.2mm iso
- put interp in config_script.yml
|
I finally decided to implement issue #91 in this PR with commit 6104cf5. @jcohenadad, i have tested this PR on a couple subjects and it works well. Could you review before merge? |
Co-authored-by: Julien Cohen-Adad <jcohen@polymtl.ca>
PaulBautin
changed the title
|
Last commits allow user to resample images to desired resolution. This new parameter must be configured in config_script.yml before running program. |
It was noticed in issue #83 that the CSA of rescaled t1w images was often overestimated. This could be caused by an increasing partial volume effect with tissue outside of the CSF. However, overestimation is less visible in t2w images, this is possibly due to better image resolution (0.8x0.8x0.8mm instead of 1x1x1mm). By resampling the T2w data to 1mm at the very beginning, it would be possible to see if difference between t1w and t2w results is caused by difference of native resolution.
DONE:
FIX #91