v0.5

This release contains the files and the trained nnUNet model to predict WM and GM in mouse spinal cord. We updated the code and instructions to use the Spinal Cord Toolbox (SCT) v6.2 instead of using a python inference script (to use the python inference script instead of SCT refer to release v0.4).

⚠️ The model is the same as in v0.4 (hence it is not added in the assets).

What's Changed

• Cleaning repo with new SCT command by @plbenveniste in #46

Full Changelog: v0.4...v0.5

v0.4

This release contains the files and the trained nnUNet model to predict WM and GM in mouse spinal cord. We updated the code to function with nnUNet 2.2.1 and added a script to correct the file's header with the correct resolution.

What's Changed

• Simplification of README by @plbenveniste in #42
• Updating inference script for nnunet 2.2.1 by @plbenveniste in #43

Full Changelog: v0.3...v0.4

Please download the 3D nnU-Net model in the "Assets" below 👇 and unzip before running test.py.

v0.3

What's Changed

• Compute uncertainty via standard deviation by @jcohenadad in #34
• Training of a nn-Unet-v2 by @plbenveniste in #32

New Contributors

• @plbenveniste made their first contribution in #32

Full Changelog: v0.2...v0.3

Please download the 3D nnU-Net model in the "Assets" below 👇 and unzip before running test.py.

v0.2

• Add ground truth for mice 10-->145 (ie: all are now included)
• Consider using ensemble/bagging for higher performance on prediction #27
• Add citation for this repos #26
• Compare with/without data augmentation applied to validation set #13
• Try smoothing (esp. along z) to gain performance at test-time #29

Full Changelog: v0.1...v0.2

Please download models in the "Assets" below 👇 and unzip before running test.py.

v0.1

What's Changed

• Introduce MONAI-compatible training script by @jcohenadad in #1

New Contributors

• @jcohenadad made their first contribution in #1

Full Changelog: https://github.com/ivadomed/model_seg_mouse-sc_wm-gm_t1/commits/v0.1