FIESTA Flow is a nextflow pipeline for white matter bundles recognition in tractography.
To use this pipeline you must have access to 6 documents - namely:
- best_model_contrastive_tractoinferno_hcp.pt
- rbx_atlas_v10.json
- mni_masked.nii.gz
- rbx_atlas (folder containing WM atlas bundles, AC.trk, AF_L.tr, ...)
- thresholds_contrastive_tractoinferno_hcp.json
- number_rejection_sampling.json
- max_total_sampling.json
- ratio.json
- degree.json
- white_matter_mask.json
- docker image felixdumais1/tractolearn-docker:
<version>
- Nextflow
- Singularity/Docker
To launch the pipeline, replace <version> by the current fiesta version, and run:
nextflow run main.nf \
--input inputs \
--model /path/to/best_model_contrastive_tractoinferno_hcp.pt \
--atlas_config /path/to/rbx_atlas_v10.json \
--atlas_anat /path/to/mni_masked.nii.gz \
--atlas_directory /path/to/data/atlas/pop_average \
--atlas_thresholds /path/to/thresholds_contrastive_tractoinferno_hcp.json \
--device cuda \
--number_rejection_sampling_config /path/to/number_rejection_sampling.json \
--max_total_sampling_config /path/to/max_total_sampling.json \
--batch_sampling 5000 \
--ratio_atlas_bundle_config /path/to/ratio.json \
--register_processes 24 \
--processes 24 \
--degree_config /path/to/degree.json \
--output_dir results \
--white_matter_config /path/to/white_matter_mask.json \
--fa_threshold 0.1 \
-resume \
-with-singularity felixdumais1/fiesta:<version>
See USAGE or run nextflow run main.nf --help
Please refer to the tractolearn "How to cite"
section to know how to cite this work.
Please refer to the tractolearn "Patent"
section to know about the patent related to this work.
This software is distributed under a particular license. Please see the LICENSE file for details.