deeds

deedsBCV with its defaults (MIND-SSC, discrete optimisation), from the corpus's rigidly aligned pair, on an isotropic grid as its README asks for thick slices.

RegistrationCPUversion 00b4653every datasetMIT

Get the app Source

Run it

With its KonfAI app, at the benchmark's settings (the fidopt app repositories are private for now: a Hugging Face token with access opens them):

pip install konfai-apps
konfai-apps infer fideus-labs/fidopt-registration:Deeds -i fixed.nii.gz -i moving.nii.gz -o out

The app on Hugging Face: its model, its settings and how it is scored.

Where it works

Its place on each kind of problem it has results on, among the methods ranked there. A method that leads one problem can trail another: read the one closest to yours.

Problem	Place	Ranked	Datasets
Repeat scans, same patient	1	of 5	1 of 1
Simulated thoracic CBCT, known motion	2	of 4	1 of 1
Repeat chest CT, same patient, tumour truth	5	of 5	1 of 1
Two modalities, same patient		not ranked yet	2 of 5
Same patient, truth made by a program		not ranked yet	2 of 8

Results

Dataset	#	Read by	Score	Change	SDlogJ	Failed	Time s
PROMISE12	1	Dice	0.814	+0.198	0.086	0	18
PSMA-PET-CT-Lesions follow-up (Learn2Reg 2026 PSMAReg)	1	TRE mm	5.13	−2.45	0.088	0	262
Pelvic-Reference-Data	4	TRE mm	2.40	−0.32	0.186	0	74
Pelvic landmarks CT/CT (LUMC)	1	Dice	0.758	+0.143	0.090	0	129
Pelvic landmarks MR/CT (LUMC)	1	Dice	0.681	+0.152	0.081	0	24
Cardiac CT of growing pigs (A heart for pigs)		Dice	0.714	=	0.000	10
Prostate-MRI-US-Biopsy	2	Dice	0.514	−0.105	0.155	0	15
LCTSC CT and simulated CBCT, known deformation	2	TRE mm	0.29	−0.95	0.018	0	21
QIN-PROSTATE-Repeatability	1	Dice	0.772	+0.128	0.066	0	25
QUADRA_HC, whole-body CT test-retest	2	Dice	0.843	+0.291	0.085	0	195
RIDER Lung CT	5	Dice	0.780	+0.044	0.070	0	260

Source: github.com/mattiaspaul/deedsBCV
Reference: Heinrich et al., IEEE Transactions on Medical Imaging 2013