r20240610

What's Changed

• Updated README file with reproducibility steps
• New model based on nnUnet #75
• Comparison of previous model_seg_ms_lesion_mp2rage and this model #81
• List of subjects used for training -> added to the datasplits folder
• Version of the dataset used to train -> added to the dataset_split.md file
• Update README file for use this model with Slicer
• Model .zip file friendly with Slicer and sct_deepseg

r20230925

Zenodo release

r20230310

What's Changed

• Update pre-processing script to use new SCT dilate feature by @jcohenadad in #71
• New model. See #49
• Edited the JSON file to change the default prediction suffix to _lesion

Full Changelog: r20230210...r20230310

r20230210

What's Changed

• Updated README with reproducibility steps and latest syntax by @uzaymacar in #31
• Added image for model overview to README by @uzaymacar in #35
• Removed comments for args for task and centerline by @uzaymacar in #37
• Script for benchmarking of SC and lesion segmentation by @uzaymacar in #40
• New config file compatible with ivadomed latest version by @jcohenadad in #51
• Config file for r20230210 by @jcohenadad in #61
• Introducing Ensembling: #63

Warning
Cropping based on spinal cord segmentation should be exactly the same as in the preprocessing. In this case:
https://github.com/ivadomed/model_seg_ms_mp2rage/blob/54d0d00e87cd3901f2ed23ce8df349d18623e8ec/preprocessing/preprocess_data.sh#L122-L124.
Also note that the dilation of the mask can now be replaced with the -dilate feature in sct_crop_image, see: spinalcordtoolbox/spinalcordtoolbox#4028

New Contributors

• @jcohenadad made their first contribution in #51

Full Changelog: r20211223...r20230210

r20211223

What's Changed

• Aggregation and Cropping in Preprocessing by @uzaymacar in #5
• Added Lesion QC for Preprocessing by @uzaymacar in #12
• Clean Data Processed Folder in Preprocessing by @uzaymacar in #7
• Documentation for centerline method params by @uzaymacar in #17
• Preparing data in preprocessing for SC segmentation task by @uzaymacar in #18
• Unified generation of clean repositories for SC seg and lesion seg by @uzaymacar in #24
• ivadomed configs for training spinal cord segmentation by @uzaymacar in #19
• ivadomed configs for training lesion segmentation by @uzaymacar in #11

Full Changelog: https://github.com/ivadomed/model_seg_ms_mp2rage/commits/r20211223