fidoptSpine-generic, spinal cord
Sagittal 3D T2-weighted MR of the cervical spine of one adult, with the spinal cord segmented by a network and corrected by hand.
Read before comparing (2)
- Cases. Both images of each of SpineGeneric's pairs, as delivered: a test pair's images are test cases, a development pair's development cases.
- Images. Sagittal 3D acquisitions at 0.8 mm, 56 to 64 slices across by 320 by 320, from the posterior fossa to the upper thoracic spine, defaced by the provider, as 64-bit floats. Values differ by site: up to 1700 to 4100 on most scanners, up to 1.8 million on Sherbrooke's Philips scanner. No mask is given.
| Accuracy | Cost | |||||||
|---|---|---|---|---|---|---|---|---|
| # | Method | Dice | Surface Dice | HD95 mm | Time s | RAM GB | GPU GB | Images |
| 1 | KonfAI segmentation app · TotalSegmentator (total_mr), the authors' packageGPUapp | 0.513 ±0.018 | 0.145 ±0.024 | 6.74 ±2.73 | 14.0 | 3.1 | ||
| 2 | KonfAI segmentation app · TotalSegmentator (total_mr), KonfAI appGPUapp1 failed | 0.482 ±0.046 | 0.124 ±0.020 | 6.42 ±1.82 | 2.1 | 5.2 | ||
| 3 | KonfAI segmentation app · IMPACTSeg body, KonfAI appGPUapp | 0.000 ±0.000 | 0.000 ±0.000 | 199.19 ±30.06 | 8 | 1.8 | 3.8 | |
The order averages each method's ranks on Dice, Surface Dice, HD95 mm, so the first one is not always the best in every column. ± is the 95% interval of the mean over the 24 cases. Two neighbours are tied when a paired Wilcoxon test on the shared cases does not find the upper one better on Dice (5%): the places a method could hold are then shown under its own. The best value of each column is in colour. Click a row to see its images, a column title to sort by it. Time and memory measured on Laptop.
Dice: how much the organ contours overlap, 1 is perfect. Surface Dice: the share of the contours' surfaces within a tolerance of each other, 1 is perfect. HD95: distance between the contours' surfaces.
Case by case
Click a score to look at that result: the image, the truth in green and the method's map in red. Underlined scores have their image; for the others you see the case without the result.
About the data
- Truth
- contours by a program, verified by people the spinal cord mask of each image (label 1): a network's output (sct_deepseg_sc) corrected by hand
- Source
- github.com/spine-generic/data-multi-subject/tree/r20250310
doi.org/10.5281/zenodo.4299140
doi.org/10.1038/s41597-021-00941-8
doi.org/10.1038/s41596-021-00588-0
spine-generic.readthedocs.io/en/latest/analysis-pipeline.html
github.com/spine-generic/data-multi-subject/pull/178
spinalcordtoolbox.com/stable/overview/concepts/labeling-conventions.html - Licence
- CC BY 4.0
22 cases for tuning, never tested
sub-tokyoSkyra02_sub-tokyoSkyra01/fixed, sub-tokyoSkyra02_sub-tokyoSkyra01/moving, sub-sherbrooke01_sub-sherbrooke02/fixed, sub-sherbrooke01_sub-sherbrooke02/moving, sub-tokyo750w06_sub-tokyo750w04/fixed, sub-tokyo750w06_sub-tokyo750w04/moving, sub-tokyoSkyra06_sub-tokyoSkyra04/fixed, sub-tokyoSkyra06_sub-tokyoSkyra04/moving, sub-hamburg06_sub-hamburg04/fixed, sub-hamburg06_sub-hamburg04/moving, sub-ucdavis04_sub-ucdavis01/fixed, sub-ucdavis04_sub-ucdavis01/moving, sub-cardiff01_sub-cardiff06/fixed, sub-cardiff01_sub-cardiff06/moving, sub-cmrrb01_sub-cmrrb02/fixed, sub-cmrrb01_sub-cmrrb02/moving, sub-geneva01_sub-geneva04/fixed, sub-geneva01_sub-geneva04/moving, sub-nottwil05_sub-nottwil04/fixed, sub-nottwil05_sub-nottwil04/moving, sub-beijingGE04_sub-beijingGE03/fixed, sub-beijingGE04_sub-beijingGE03/moving