fidoptSynthRAD2023 MR/CT, brain
CT (fixed) and MR (moving) of one patient's brain, rigidly aligned as the challenge delivers them (SynthRAD2023, Task 1).
Read before comparing (4)
- Labels. TotalSegmentator through its KonfAI app at a pinned revision, one recipe for every case (truth.labels_about, ../synthrad.py). A method that registers on that network's features is scored against its own view of the anatomy.
- Specks. The network also marks structures a head does not hold, each on one image of a case: a humerus, scapula or femur on 7 test MRs (2 to 2,506 voxels of 1 mm³), a gluteus on one test CT (30 voxels), a scapula on two development MRs. They are kept as segmented and never scored, since a structure is scored only where both images hold it: every case is scored on its spinal cord and brain.
- Mask. The challenge's body mask, the same for both images.
- Baseline. The pair is used rigidly aligned, as delivered. With no registration the 18 test pairs score a label Dice of 0.57 (0.45 to 0.68): most methods score below the unregistered pair here.
No method has a result on every pair of this dataset yet.
| Accuracy | Smoothness | Cost | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| # | Method | Dice | Change | HD95 mm | SDlogJ | Folds % | Time s | RAM GB | GPU GB | Images |
| No registration | 0.574 ±0.033 | 9.00 ±1.62 | 0.000 | 0.00 | ||||||
2 results not ranked: an older run, pairs still missing, or a failure on every pair
| Accuracy | Smoothness | Cost | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| # | Method | Dice | Change | HD95 mm | SDlogJ | Folds % | Time s | RAM GB | GPU GB | Machine | Images |
| ConvexAdam (MIND)GPUapp8/18 pairs | 0.635 ±0.060 | +0.061 | 7.60 ±2.49 | 0.075 | 0.49 | 6 | 2.6 | 3.0 | Laptop | ||
| NiftyRegCPUapp8/18 pairs | 0.496 ±0.048 | −0.078 | 15.88 ±5.39 | 0.074 | 0.19 | 45 | 1.1 | Laptop | |||
Dice: how much the organ contours overlap, 1 is perfect. HD95: distance between the contours' surfaces. SDlogJ: how irregular the deformation is, 0 is perfectly smooth. Folds: share of the image where the deformation folds over itself.
Pair by pair
Click a score to look at that result: the fixed image, and over it the moving image after registration, with the annotations. Underlined scores have their image; for the others you see the pair without the result. A green cell gained against no registration on that pair, a red one lost.
About the data
- Test pairs
- 18 test and 16 development patients among the 180 training patients, chosen before any result, centres 1BA, 1BB and 1BC alternated. The test list takes each centre's consecutive patients in ascending order from 1BA014, 1BB066 and 1BC001, development patients skipped. The development list holds 1BA's first three, seven others and each centre's last two in descending order. No patient is in both.
- Truth
- contours by a program TotalSegmentator through its KonfAI app (VBoussot/TotalSegmentator-KonfAI, revision 07185697, konfai-apps of KonfAI 1.8.7), run by us on each image with one recipe (synthrad.py): task total on the CT, task total_mr on the MR, both reduced to the structures the two tasks name, in total_mr's ids
- Licence
- CC BY-NC 4.0
16 pairs for tuning, never tested
brain_1BA001, brain_1BA005, brain_1BA012, brain_1BA103, brain_1BB076, brain_1BB033, brain_1BB090, brain_1BC046, brain_1BC008, brain_1BC039, brain_1BA379, brain_1BB205, brain_1BC094, brain_1BA368, brain_1BB200, brain_1BC090