Skip to main content
. 2021 Dec;11(12):4895–4916. doi: 10.21037/qims-21-175

Table 2. Overview of deep iterative registration studies.

Study Learning Transform Modality ROI Public dataset for training/validation Evaluation metrics
(76) Metric Deformable MRI-MRI Brain LONI LPBA40 DSC
(77) Metric Deformable MRI-MRI Brain IXI, ALBERTs DSC
(78) Metric Deformable MRI-MRI Brain IXI FRE, DSC
(79) Metric Rigid MRI-US Prostate PROMISE12, Prostate-3T TRE
(80) Metric Deformable MRI-MRI Brain OASIS, ABIDE DSC
(81) Metric Rigid & Deformable MRI-MRI Brain Simulated brain DSC
(82) Metric Deformable MRI-MRI Brain LPBA40, CUMC12, MGH10, IBSR18 DSC, Jaco. Det.
(83) RL agent Rigid CT-CBCT Spine and heart N/A TRE
(84) RL agent Deformable CT-CT Thorax, abdomen, and pelvis N/A Euclidean distance
(85) RL agent Deformable MRI-MRI Prostate PROMISE12, Prostate-3T DSC, HD
(86) RL agent Affine MR-CT Brain N/A TRE

CBCT, cone-beam computed tomography; CT, computed tomography; DSC, Dice coefficient; FRE, fiducial registration error; HD, Hausdorff distance; MRI, magnetic resonance imaging; RL, reinforcement learning; ROI, region of interest; TRE, target registration error.