cobralab / minc-toolkit-extras Goto Github PK

antsRegistration_affine_SyN

• specify final iterations
• specify multispectral metrics and weights
• allow different image pairs and metrics for affine vs syn
• handling resampling of multiple image pairs properly
• allow multiple initializations

ants_generate_iterations

• specify multispectral metrics and weights

Or at least throw an error.

I am running into issues with create_civet_image.sh failing to parse column names with commas in them. The 'left' file below ends up being empty, as is the 'right' file in the same folder.

bc might not be available

Multispectral registration fails because the weight array isn't properly specified

@gdevenyi,

First, thanks for all of the exceptionally cool improvements you make to existing programs. ants_generate_iterations.sh and antsRegistration_affine_SyN.sh are fantastic and have dramatically improved not only my mouse templates but also the registration of those templates to standard spaces.

Two things:

1. If anything ultimately comes out of using these (e.g., manuscripts), let me know how you'd like these scripts credited.
2. If I just need a rigid registration between a _b_0 image (diffusion) and a T2 in order to move my diffusion data to T2 space, what are the right flags to set for antsRegistration_affine_SyN.sh. Is it as simple as antsRegistration_affine_SyN.sh $moving $fixed dwi2struct_ --skip-nonlinear --linear-type rigid?

Need to fix paths for supporting datafiles.

Could you please share the python environment required to run the t2star_fit.py and t2star_fit_simpleitk.py functions?

I'm running into several issues when trying to install the imported packages one by one.

Thanks.

Hi @gdevenyi

Running antsRegistration_affine_SyN.sh in a Singularity image. Installed from current Github as of last week. I get the following after stage 3 (AffineTransformation):

antsRegistration_affine_SyN.sh $TPL ${initmask}/mean_bzero.nii.gz ${initmask}/tpl2dwi_ --fast --moving ${initmask}/tpl_lap_norm.nii.gz --fixed ${initmask}/mean_bzero_lap_norm.nii.gz
...
/opt/minc-toolkit-extras/antsRegistration_affine_SyN.sh: line 860: _arg_weights[$((i+1))]: unbound variable

All files exist. I can run registration between the TPL and mean_bzero as well as *lap_norm files independently, just not as multispectral.

Any thoughts? Is it because weights isn't specified (I just thought of that as I wrote it...)

Command run :

./create_morpho_image.sh new_lefthc.obj ../lm/tvalue_CA_left_icv.csv 2.17406767435032,2.83736873702781 new_righthc.obj ../lm/tvalue_CA_right_icv.csv 2.28660590804178,2.96056946858748 tvalue-groupcatCTRL ../morpho_images/lm_CA_icv.png &> create_morpho_image.log

I'm getting the thresholds automatically, that's why the long format. Using "normal" format does not change the issue.

The two csv files have the tvalue-groupcatCTRL column.

Content of create_morpho_image.log :

Value range: -10.3427 9.13157
Value range: -10.3427 9.13157
Value range: -8.1667 6.80638
Value range: -8.1667 6.80638
Ray Tracing: ..................................................................
Ray Tracing: ..................................................................
Ray Tracing: ..................................................................
Ray Tracing: ..................................................................
Ray Tracing: ..................................................................
Ray Tracing: ..................................................................
Ray Tracing: ..................................................................
Ray Tracing: ..................................................................
/tmp/tmp.HtXGFYoNLc

Content of TMPDIR :

1.png
2.png
3.png
4.png
back.png
back.rgb
bottom.png
bottom.rgb
coldmetalinv.png
front.png
front.rgb
left_comb.obj
left_lateral.png
left_lateral.rgb
left_medial.png
left_medial.rgb
left_pos.obj
redmetalinv.png
right_comb.obj
right_lateral.png
right_lateral.rgb
right_medial.png
right_medial.rgb
right_pos.obj
rmincblue.map
rmincred.map
top.png
top.rgb
tvalue_CA_left_icv.csv
tvalue_CA_right_icv.csv
unlabelled.png

lm_CA_icv.png

unlabelled.png

Print a warning if both images are not masked, but there's no reason we can't extract only the moving or only the fixed if provided.

This comes up for things like CBF->T1 and similar, where one of the images has no skull signal intrinsicly

Hi,

I'm sorry for asking , in the command instruction :

antsRegistration_affine_SyN.sh  [...]\
 --fixed multi_f1 --moving multi_m1\
 --fixed multi_f2 --moving multi_m2 \
--weights 0.1,0.2, 0.7  [....]\
 <movingfile_m0> <fixedfile_f0> <outputbasename>

A single value, which disables weighting, or a comma separated list of weights, ordered primary pair, followed by multispectral pairs

so if I define --weights 0.1,0.2,0.7

movingfile_m0 fixedfile_f0 are the primary pair, weight will be 0.1

and [--fixed <multi_f1] [--moving multi_m1] are the first multispectral pairs, weight will be 0.2
and [--fixed <multi_f2] [--moving multi_m2] are the first multispectral pairs, weight will be 0.7

Is it true or the opposite ?
Thanks in advance,
Valéry

• allow for control of colourbar labels
• allow for changing colourmaps
• allow for specifying image title
• integrate proper command line parsing
• allow for single-direction maps (no bicolour)
• provide command line for multiple morpho objects and maps
• orientations to generate
• orientation labels on/off
• background colour
• check length of inputs correspond properly to object being used.

Hello! I am hoping to use ants_generate_iterations.py to construct a more robust group template, but I am unsure what the --min and --max arguments are referring to. Are they referring to dimensions of the input images, or the dimensions of the voxels in the input images? Or something else?