fidoptSynthRAD2023, MR to CT, pelvis
Pelvic MR of one patient and the CT to synthesise from it, the CT brought onto the MR's grid by the challenge (SynthRAD2023, Task 1).
Read before comparing (6)
- Case. The challenge's pair as delivered, copied byte for byte from Task1.zip: the MR, the CT the organisers registered onto its grid and their body mask.
- Mask. The challenge's body mask: the synthetic image is scored inside it, as the challenge scores it.
- Alignment. How the organisers registered the CT to the MR bounds how far a perfect synthesis can match it: what that registration left misaligned counts as error for every method.
- Baseline. Inside the mask on the 14 test cases, with KonfAI's criteria at CT's range of 4095, a constant at the target's median scores an MAE of 258 (211 to 304) HU, a PSNR of 19.7 (18.9 to 20.3) dB and an SSIM of 0.23 (0.10 to 0.49).
- Metal. pelvis_1PA155, a test case, holds a hip prosthesis: a void on the MR, metal on the CT.
- Overlap. No other descriptor names SynthRAD2023 or its record. The provider says its two tasks 'do not necessarily contain the same patients', so each drawn case's CT was registered rigidly to every CT of SynthRAD_CBCTCT (34) and SynthRAD_CBCTCT_Syn and correlated inside the bodies: at most 0.88, where one CT on two grids reads 0.98 or more.
No method has a result on every case of this dataset yet.
Under a licence that is not open
Methods whose code or weights are released for non-commercial use, or under share-alike or custom terms: shown for reference, ranked nowhere.
| Accuracy | Cost | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| # | Method | MAE | PSNR dB | SSIM | Anatomy Dice | Anatomy HD95 mm | Time s | RAM GB | GPU GB | Machine | Images |
| nnsyn · MR to CT, abdomenGPUapptrained on CC-BY-NC-4.0 dataolder run | 92.9 ±13.1 | 26.00 ±1.12 | 0.721 ±0.037 | 0.693 ±0.072 | 12.83 ±4.68 | 32 | 4.5 | 2.7 | Laptop | ||
MAE: mean absolute error of a synthetic image against its target, inside the body, in the image's units. PSNR: the same by signal-to-noise ratio, higher is better. SSIM: the same by structural similarity, 1 is perfect. Anatomy Dice and HD95: TotalSegmentator run on the synthetic and on the target image, the structures compared; shown, not ranked.
Case by case
| MAE | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 |
|---|
Click a score to look at that result: the target image, and over it the method's synthetic image. Underlined scores have their image; for the others you see the case without the result.
About the data
- Test cases
- 14 test patients drawn with seed 2026 by centre (1PA, 1PC) in proportion, then 13 development patients, the last of the same seeded order, the centres in proportion (9 1PA, 4 1PC). The pool is 146 of SynthRAD2023 Task 1's 180 pelvis training patients, once SynthRAD_MRCT's 18 test and 16 development patients are out, since its public fixed image is this target.
- Truth
- target image as released the patient's CT as the challenge delivers it, registered to the MR by the organisers, in Hounsfield units
- Licence
- CC BY-NC 4.0
13 cases for tuning, never tested
pelvis_1PA098, pelvis_1PC071, pelvis_1PA112, pelvis_1PA056, pelvis_1PA161, pelvis_1PC024, pelvis_1PA026, pelvis_1PA111, pelvis_1PC095, pelvis_1PA094, pelvis_1PA108, pelvis_1PC042, pelvis_1PA035