Releases: spinalcordtoolbox/spinalcordtoolbox
6.4
Note
Version 6.4 of SCT was originally released on Aug 8, 2024. However, the release was then updated on Oct 2, 2024 to backport 2 fixes for sct_deepseg
related bugs.
Notable changes include:
- Feature: Add midsagittal tissue bridges to
sct_analyze_lesion
. View pull request - Feature: Support custom labels from
info_label.txt
forsct_analyze_lesion -f
. View pull request - Feature: Add slicewise analysis for
sct_analyze_lesion -f
. View pull request - Feature: Track
sct_deepseg
model provenance withsource.json
(in model folder) and JSON sidecar (in output). View pull request - Feature: Add T2w dog template to
sct_download_data
. View pull request - Enhancement: Update contrast agnostic
sct_deepseg
model to v2.4 (now improved for lumbar t2w + PSIR/STIR images). View pull request - Enhancement: Update SCI
sct_deepseg
model to SCIsegV2. View pull request - Enhancement: Improve
sct_deepseg
rootlets seg QC report by improving the cropping, centering, and colormap. View pull request - Enhancement: Use
LazyLoader
class to minimize startup time for all CLI scripts. View pull request - Bug: Fix straightening transformations for images with "tilted" qform/sform. View pull request
Full release notes and Changelog
Results of batch_processing.sh
on Ubuntu 20.04
~~~
Version: git-master-44471c6bfb2d9c9bb507dd6a99120db4508db84a
Ran on: Linux fv-az2030-237 5.15.0-1068-azure
Duration: 0hrs 15min 6sec
---
t2/csa_c2c3.csv: 73.90440963005199 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89183952519573 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530312082996 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.789823839063319 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.44251331713387 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.43643151888196 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7788101713862418 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7765826899727847 [Row 2, WA()]
~~~
Results of batch_processing.sh
on macOS 12 (Monterey)
~~~
Version: git-master-44471c6bfb2d9c9bb507dd6a99120db4508db84a
Ran on: Darwin Mac-1723200118668.local 21.6.0
Duration: 0hrs 19min 51sec
---
t2/csa_c2c3.csv: 73.90440963005199 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89183952519573 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530312082996 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.789823839063319 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.44251331713387 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.424413121632384 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.779190298495622 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7762361654476364 [Row 2, WA()]
~~~
Results of batch_processing.sh
on Windows 2019
~~~
Version: git-master-44471c6bfb2d9c9bb507dd6a99120db4508db84a
Ran on: MINGW64_NT-10.0-20348 fv-az534-455 3.4.10-87d57229.x86_64
Duration: 0hrs 18min 59sec
---
t2/csa_c2c3.csv: 73.90440963005199 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89183952519573 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530312082996 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.789823839063319 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.44251331713387 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.43643151888196 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7788101713862418 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7765826899727847 [Row 2, WA()]
~~~
6.3
Notable changes include:
- Feature: Add QC report for
sct_deepseg
. View pull request - Feature: Add CanProCo-based MS lesion segmentation model to
sct_deepseg
. View pull request - Feature: Add EPI-BOLD fMRI spinal cord segmentation model to
sct_deepseg
. View pull request - Enhancement: Update
contrast-agnostic
SC segmentation model (sct_deepseg
) to the latest version (v2.3). View pull request - Enhancement: Switch to using mean magnitude for output
moco_params.tsv
file used for QC. View pull request - Documentation: Add links to new 2024 SCT review paper to prominent locations in documentation. View pull request
Full release notes and Changelog
Results of batch_processing.sh
on Ubuntu 20.04
~~~
Version: git-HEAD-f0e3766f4d663d28fbb6b718cd0f76bd203a0971
Ran on: Linux fv-az1533-44 5.15.0-1061-azure
Duration: 0hrs 16min 46sec
---
t2/csa_c2c3.csv: 73.87680055661444 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89183952519573 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530312082996 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.515230379832953 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93314479093739 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.417007086785446 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7758118462499376 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.773312472697186 [Row 2, WA()]
~~~
Results of batch_processing.sh
on macOS 12 (Monterey)
~~~
Version: git-HEAD-f0e3766f4d663d28fbb6b718cd0f76bd203a0971
Ran on: Darwin Mac-1714066641279.local 21.6.0
Duration: 0hrs 21min 32sec
---
t2/csa_c2c3.csv: 73.87095978136215 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89183952519573 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530312082996 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.515230379832953 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93314479093739 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.41827867818828 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7761900744878227 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7765149211365178 [Row 2, WA()]
~~~
Results of batch_processing.sh
on Windows 2022
~~~
Version: git-HEAD-f0e3766f4d663d28fbb6b718cd0f76bd203a0971
Ran on: MINGW64_NT-10.0-20348 fv-az1105-632 3.4.10-87d57229.x86_64
Duration: 0hrs 22min 13sec
---
t2/csa_c2c3.csv: 73.87680055661444 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89183952519573 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530312082996 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.515230379832953 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93314479093739 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.417007086785446 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7758118462499376 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.773312472697186 [Row 2, WA()]
~~~
6.2
Notable changes include:
- Feature: Integrate 4 new nnUNet/MONAI models into
sct_deepseg
(contrast-agnostic softseg, SCI lesion/SC seg, spinal nerve rootlet seg, mouse GM/WM seg). View pull request - Feature: Save QC records to browser local storage to avoid losing data on refresh. View pull request
- Feature: Update PAM50 template to include new
PAM50_rootlets.nii.gz
file. View pull request - Bugfix: Fix straightening error during registration if 3+ labels are supplied and topmost disc label is not C1. View pull request
- Bugfix: Mitigate scaling issues (
1.0
->0.999
) due to float/int datatype mismatches between header and array. View pull request - Installation: Specify Rosetta 2 as a requirement for installation on Apple silicon (M1, M2, M3). View pull request
- Documentation: Port changes from SCT Course 2023 Google Slides to the web tutorials. View pull request
Full release notes and Changelog
Results of batch_processing.sh
on Ubuntu 20.04
~~~
Version: git-master-6962e03b5906ab3466e6e330438dbea58d949407
Ran on: Linux fv-az1018-974 5.15.0-1054-azure
Duration: 0hrs 17min 15sec
---
t2/csa_c2c3.csv: 73.8768043493825 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89184331879929 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530487782326 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.487533885581323 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93464311280606 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.417007018368906 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7758118559608612 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7733124731836604 [Row 2, WA()]
~~~
Results of batch_processing.sh
on macOS 11 (Big Sur)
~~~
Version: git-master-6962e03b5906ab3466e6e330438dbea58d949407
Ran on: Darwin Mac-1708024085684.local 20.6.0
Duration: 0hrs 29min 31sec
---
t2/csa_c2c3.csv: 73.8709635738436 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89184331879929 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530487782326 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.487533885581323 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93464311280606 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.41827830455262 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7761900951571409 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7765149187171309 [Row 2, WA()]
~~~
Results of batch_processing.sh
on Windows 2019
~~~
Version: git-master-6962e03b5906ab3466e6e330438dbea58d949407
Ran on: MINGW64_NT-10.0-17763 fv-az1488-920 3.4.9-be826601.x86_64
Duration: 0hrs 23min 38sec
---
batch_processing.sh: line 270: D:\a\spinalcordtoolbox\spinalcordtoolbox/python/envs/venv_sct/bin/python: No such file or directory
t2/csa_c2c3.csv: 73.8768043493825 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89184331879929 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.23530487782326 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.487533885581323 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93464311280606 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.417007018368906 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7758118559608612 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7733124731836604 [Row 2, WA()]
~~~
6.1
This minor release of SCT has been developed in preparation for the 2023-10-20 SCT Course. It contains a significant update to the PAM50 template, important documentation improvements, and many other bugfixes and tweaks.
Notable changes include:
- Feature: Update PAM50 template link to include cord and lumbar label changes. View pull request
- Feature: Add function to output the axial damage ratio for
sct_analyze_lesion
. View pull request - Documentation: Add tutorial for
sct_compute_compression
View pull request - Documentation: Add tutorial for lumbar segmentation and registration. View pull request
- Documentation: Update Docker installation instructions for Linux/macOS/Windows. View pull request
- Maintenance: Remove
-s
functionality fromsct_warp_template
and add a deprecation warning. View pull request - Bugfix: Fix distorted registration due to straightening bug in
get_closest_to_absolute
. View pull request - Bugfix Use pandas for
.csv
saving insct_compute_compression
to correctly merge existing output metric columns. View pull request
Full release notes and Changelog
Results of batch_processing.sh
on Ubuntu 20.04
~~~
Version: git-master-f5a46f328fe797b3d7c0e3844e17ad3f8add5ee1
Ran on: Linux fv-az619-734 5.15.0-1050-azure
Duration: 0hrs 26min 56sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.25734847835911 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.48783482885619 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088245 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.39463563966487 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793857114020448 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.775688329824498 [Row 2, WA()]
~~~
Results of batch_processing.sh
on macOS 11 (Big Sur)
~~~
Version: git-master-f5a46f328fe797b3d7c0e3844e17ad3f8add5ee1
Ran on: Darwin Mac-1699196155540.local 20.6.0
Duration: 0hrs 28min 56sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.25734847835911 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.41599937353151 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7807687336793107 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.774647306583997 [Row 2, WA()]
~~~
Results of batch_processing.sh
on Windows 2019
~~~
Version: git-master-f5a46f328fe797b3d7c0e3844e17ad3f8add5ee1
Ran on: MINGW64_NT-10.0-17763 fv-az981-219 3.4.7-25de8b84.x86_64
Duration: 0hrs 22min 58sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.25734847835911 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088248 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.39462068731342 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793857115174943 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7756883298906552 [Row 2, WA()]
~~~
SCT v6.0
This major release provides significant improvements for how SCT is installed on all platforms, as well as many new features and bugfixes.
Notable changes include:
- Installation: Allow
install_sct
to be run standalone (without downloading "Source code" archive). View pull request - Installation: Use Miniconda instead of built-in Python for Windows installations. View pull request
- Feature: Add new CLI script to compute normalized metric ratios (MSCC, etc.) for compressed levels. View pull request
- Feature: Add new
-histo
option tosct_warp_template
to warp the newly-added PAM50 histology files. View pull request - Feature: Add new sagittal mosaic option for
sct_deepseg_lesion
QC report. View pull request - Feature: Add support for model ensembles to
sct_deepseg
and use it formp2rage_lesion
model. View pull request - Feature: Add new
-project-centerline
option tosct_label_utils
to project an image on the spinal cord centerline. View pull request - Feature: Add new
-centerline-soft
option tosct_get_centerline
to output a non-binary "soft" centerline. View pull request - Bugfix: Ensure that qform/sform codes are preserved when generating
sct_deepseg_sc
segmentation. View pull request
Full release notes and Changelog
Results of batch_processing.sh
on Ubuntu 20.04
~~~
Version: git-master-908998829a2a3694fa96363b358c9b662da4ae43
Ran on: Linux fv-az205-332 5.15.0-1041-azure
Duration: 0hrs 23min 3sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.25734847835911 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.38856944000351 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793405440192693 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755611783165834 [Row 2, WA()]
~~~
Results of batch_processing.sh
on macOS 11 (Big Sur)
~~~
Version: git-master-908998829a2a3694fa96363b358c9b662da4ae43
Ran on: Darwin Mac-1689358819888.local 20.6.0
Duration: 0hrs 39min 15sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.25734847835911 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.43230073944698 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7805537210812612 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7746193990233381 [Row 2, WA()]
~~~
Results of batch_processing.sh
on Windows 2019
~~~
Version: git-master-908998829a2a3694fa96363b358c9b662da4ae43
Ran on: MINGW64_NT-10.0-17763 fv-az34-210 3.4.7-ea781829.x86_64
Duration: 0hrs 26min 44sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2/csa_pam50.csv: 34.25734847835911 [Row 39, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088248 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.38856944000351 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793405440192693 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755611783165834 [Row 2, WA()]
~~~
SCT v5.8
Note: Version 5.8 of SCT was originally released on Nov 30, 2022. However, the release was then updated on Feb 17, 2023 to backport a fix for an installation related bug (24f602b), and on Jul 19, 2023 to backport two fixes for bugs causing some SCT commands to crash (b8be5c4).
Notable changes include:
- Feature:
sct_image
: Add new-stitch
option for combining sequential image scans. View pull request - Feature:
sct_run_batch
: Add new-ignore-ses
option to prioritizesub-
directories overses-
subdirectories. View pull request - Feature:
sct_process_segmentation
: For the-perslice
option, begin outputting theDistancePMJ
metric. View pull request - Enhancement: Add readability fixes for QC reports (sagittal view scaling, label text, label colormaps). View pull request
- Enhancement:
image.py
: Update header dtype property on save/load to match the datatype of the data array. View pull request - Bugfix:
sct_run_batch
: Modify-include-list
and-exclude-list
to check against parts of a directory, too. View pull request - Bugfix:
sct_run_batch
: Allow path_output parameter to start with~
. View pull request - Installation: Upgrade SCT from Python 3.7 to Python 3.8. View pull request
- Documentation: Emphasize references to PMJ method by Bédard and Cohen-Adad. View pull request
Full release notes and Changelog
Results of batch_processing.sh
on Ubuntu 20.04
~~~
Version: git-master-71e199dc14156e895880bc8e74de8409a2d238c8
Ran on: Linux fv-az259-426 5.15.0-1023-azure
Duration: 0hrs 23min 13sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.37973940532884 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793405440192693 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755611783165834 [Row 2, WA()]
~~~
Results of batch_processing.sh
on macOS 11 (Big Sur)
~~~
Version: git-master-71e199dc14156e895880bc8e74de8409a2d238c8
Ran on: Darwin Mac-1669763228410.local 20.6.0
Duration: 0hrs 28min 36sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.424258247179175 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7805537210812612 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7746193990233381 [Row 2, WA()]
~~~
Results of batch_processing.sh
on Windows 2019
~~~
Version: git-master-71e199dc14156e895880bc8e74de8409a2d238c8
Ran on: MINGW64_NT-10.0-17763 fv-az30-853 3.3.6-341.x86_64
Duration: 0hrs 24min 52sec
---
batch_processing.sh: line 267: ./python/envs/venv_sct/bin/python: No such file or directory
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088248 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.37973940532884 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793405440192693 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755611783165834 [Row 2, WA()]
~~~
SCT v5.7
🔔 Reminder for Windows users 🔔
Since version 5.6, WSL and Docker are no longer required to install SCT on Windows! 🎉
Please visit the updated Windows installation page of SCT's documentation to try out this new installation method.
Notable changes include:
- Feature: Multimodal registration using Deep Learning methods are now integrated in SCT (
sct_register_to_template
,sct_register_multimodal
). View pull request - Tutorial: New tutorial for contrast agnostic registration. View pull request
- Bugfix:
sct_process_segmentation
,sct_extract_metric
: Combine conditions when slice number and vertebral levels are both specified. View pull request - Bugfix:
sct_propseg
: Prevent from overwriting files. View pull request - Bugfix: Correctly handle output files on different drives on Windows. View pull request
- Bugfix:
sct_maths
: Don't treat a single 4D image as a sequence of 3D images in-add
/-sub
/-mul
/-div
. View pull request - Maintenance:
sct_deepseg_sc
,sct_deepseg_gm
,sct_deepseg_lesion
: Replace Tensorflow/Keras-based inference (.h5
) with onnxruntime (.ONNX
). View pull request - Testing: The
batch_processing.sh
tests now support macOS and Windows. View pull request
Full release notes and Changelog
Results of batch_processing.sh
on Ubuntu 20.04
~~~
Version: git-master-ff99c2cc6364c2241da3203ead46026e24908657
Ran on: Linux fv-az453-981 5.15.0-1014-azure
Duration: 0hrs 22min 54sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.37973940532884 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793999418865862 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755359512006275 [Row 2, WA()]
~~~
Results of batch_processing.sh
on macOS 11 (Big Sur)
~~~
Version: git-master-ff99c2cc6364c2241da3203ead46026e24908657
Ran on: Darwin Mac-1659036569665.local 20.6.0
Duration: 0hrs 27min 49sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.424258247179175 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7806055878397453 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7746118274273426 [Row 2, WA()]
~~~
Results of batch_processing.sh
on Windows 2019
~~~
Version: git-master-ff99c2cc6364c2241da3203ead46026e24908657
Ran on: MINGW64_NT-10.0-17763 fv-az158-506 3.3.5-341.x86_64
Duration: 0hrs 26min 1sec
---
batch_processing.sh: line 267: ./python/envs/venv_sct/bin/python: No such file or directory
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088248 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.37973940532884 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793999418865862 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755359512006275 [Row 2, WA()]
~~~
SCT v5.6
🔔 Notice for Windows users 🔔
Starting from version 5.6, WSL and Docker are no longer required to install SCT on Windows! 🎉
Please visit the updated Windows installation page of SCT's documentation to try out this new installation method.
Other notable changes include:
sct_deepseg
: Add model for T2w lumbar SC segmentation. View pull requestsct_deepseg
: Add new option-list-tasks-long
to print in-depth descriptions of deepseg tasks. View pull requestsct_label_vertebrae
: Update the default label cleaning behavior. View pull request- Add support for ITK-Snap + multiple viewers to
display_viewer_syntax
. View pull request - BUGFIX:
sct_analyze_lesion
: Fix computation of estimated lesion length and diameter. View pull request - BUGFIX: Set a more permissive threshold for reading the qform. View pull request
Full release notes and Changelog
Results of batch_processing.sh
~~~
Version: git-master-c642c1e3285efbd5fdb88e19a5c30da341444927
Ran on: Linux fv-az81-867 5.4.0-1077-azure
Duration: 0hrs 18min 5sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.37973940532884 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793999418865862 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755359512006275 [Row 2, WA()]
~~~
SCT v5.5
Notable changes include:
- Added new
sct_deepseg
models for MP2RAGE spinal cord and MS lesion segmentation. View pull request - Added new
sct_deepseg
model for 7T spinal cord/gray matter multiclass segmentation. View pull request - Added new Patch2Self CLI script for dMRI denoising. View pull request
- Brought back previously-removed
sct_testing
command as a light wrapper forpytest
. View pull request - Fixed bug in
sct_compute_mtr
when run on high-valued int16 data.
Full release notes and Changelog
Results of batch_processing.sh
~~~
Version: git-HEAD-daf4ce949b79e686a3ded53601e5b5f12c4ef9fd
Ran on: Linux fv-az41-165 5.4.0-1067-azure
Duration: 0hrs 20min 23sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856178 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.37973940532884 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793999143910345 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755359385435815 [Row 2, WA()]
~~~
SCT v5.4
Notable changes include:
- Feature: Measure CSA based on distance from pontomedullary junction (PMJ) using
sct_process_segmentation -pmj
. - Feature: Normalize CSA based on sex, brain volume, and thalamus volume using
sct_process_segmentaiton -normalize
. - Feature: Visualize b-vectors for multi-shell acquisitions using
sct_dmri_display_bvecs -bval
. - Feature: Compute SNR within a single 3D volume using
sct_compute_snr -method single
. - Feature: It is now possible to loop across "ses-" entities in
sct_run_batch
. - Enhancement: In
sct_detect_pmj
, the method used to determine the R-L placement of the PMJ label has been improved. - Bug: The SCT FSLeyes script is now compatible with FSLeyes v1.X.
Full release notes and Changelog
Results of batch_processing.sh
:
~~~
Version: git-master-5e54502852b4212dba2bedf3aa5de740b5660516
Ran on: Linux fv-az76-747 5.4.0-1056-azure
Duration: 0hrs 25min 53sec
---
t2/csa_c2c3.csv: 73.87711295363036 [Row 1, MEAN(area)]
t2/csa_pmj.csv: 73.89298021190447 [Row 1, MEAN(area)]
t2s/csa_gm.csv: 12.487834828856176 [Row 4, MEAN(area)]
t2s/csa_wm.csv: 64.93830702088249 [Row 4, MEAN(area)]
mt/mtr_in_wm.csv: 54.379739375142336 [Row 1, MAP()]
dmri/fa_in_cst.csv: 0.7793999143910345 [Row 1, WA()]
dmri/fa_in_cst.csv: 0.7755359385435815 [Row 2, WA()]
~~~