# 3D registration on DirLab landmarks 3,000 expert-annotated (test)

Source: [3D registration on DirLab landmarks 3,000 expert-annotated (test)](<https://www.sota2.com/research/sota/3d-registration-on-dirlab-landmarks-3-000-expert-annotated-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:35:25.250Z

Evaluation of 3D registration performance using expert-annotated landmarks from the DirLab 3,000 dataset (test set), reporting error statistics in millimeters.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | 3D registration |
| Dataset (including any stated split) | DirLab landmarks 3,000 expert-annotated (test) |
| Metric | Average Error (mm) |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 12 of 12 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Average Error (mm)](<https://www.sota2.com/research/sota/3d-registration-on-dirlab-landmarks-3-000-expert-annotated-test.md?metric=Average%20Error%20(mm)>): Lower is better; unit: SCALAR.

- [25th Percentile Error (mm)](<https://www.sota2.com/research/sota/3d-registration-on-dirlab-landmarks-3-000-expert-annotated-test.md?metric=25th%20Percentile%20Error%20(mm)>): Lower is better; unit: SCALAR.

- [Median Error (mm)](<https://www.sota2.com/research/sota/3d-registration-on-dirlab-landmarks-3-000-expert-annotated-test.md?metric=Median%20Error%20(mm)>): Lower is better; unit: SCALAR.

- [75th Percentile Error (mm)](<https://www.sota2.com/research/sota/3d-registration-on-dirlab-landmarks-3-000-expert-annotated-test.md?metric=75th%20Percentile%20Error%20(mm)>): Lower is better; unit: SCALAR.

- [Time (s)](<https://www.sota2.com/research/sota/3d-registration-on-dirlab-landmarks-3-000-expert-annotated-test.md?metric=Time%20(s)>): Lower is better; unit: TIME.

## Recorded results

| Method | Average Error (mm) | Result date | Report paper |
| --- | --- | --- | --- |
| D-RobOT (LDDMM) | 2.86 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| D-RobOT (spline) | 2.95 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| D-RobOT (raw) | 3.4 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| DispEmd+ | 3.42 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| DGCNN-CPD+ | 4.3 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| S-RobOT (LDDMM) | 5.48 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| S-RobOT (spline) | 5.72 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| CPD (non-rigid) | 9.3 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| RobOT (raw) | 9.41 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| RobOT (affine) | 10.45 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| ICP (affine) | 15.05 | Not recorded | 68fa6b2f1e209131cea2cad2 |
| Input data | 23.3 | Not recorded | 68fa6b2f1e209131cea2cad2 |

### D-RobOT (LDDMM)

Average Error (mm): 2.86. Result ID: 698e6b80ffc3901de09d7359.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Implementation](<https://github.com/uncbiag/robot>)

Recorded evaluation configuration: {"Training regime":"Supervised","Input modality":"Point clouds","Regularization":"LDDMM"}. This record does not assert independent reproduction.

### D-RobOT (spline)

Average Error (mm): 2.95. Result ID: 698e6b80ffc3901de09d7358.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Implementation](<https://github.com/uncbiag/robot>)

Recorded evaluation configuration: {"Training regime":"Supervised","Input modality":"Point clouds","Regularization":"Spline"}. This record does not assert independent reproduction.

### D-RobOT (raw)

Average Error (mm): 3.4. Result ID: 698e6b80ffc3901de09d7357.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Implementation](<https://github.com/uncbiag/robot>)

Recorded evaluation configuration: {"Training regime":"Supervised","Input modality":"Point clouds","Variation":"Raw"}. This record does not assert independent reproduction.

### DispEmd+

Average Error (mm): 3.42. Result ID: 698e6b80ffc3901de09d7354.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67951e209131cea172d0.md>)

Recorded evaluation configuration: {"Training regime":"Supervised","Input modality":"Image keypoints"}. This record does not assert independent reproduction.

### DGCNN-CPD+

Average Error (mm): 4.3. Result ID: 698e6b80ffc3901de09d7353.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66631e209131cea10058.md>)

Recorded evaluation configuration: {"Training regime":"Supervised","Input modality":"Image keypoints"}. This record does not assert independent reproduction.

### S-RobOT (LDDMM)

Average Error (mm): 5.48. Result ID: 698e6b80ffc3901de09d7356.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Implementation](<https://github.com/uncbiag/robot>)

Recorded evaluation configuration: {"Training regime":"Supervised","Input modality":"Point clouds","Regularization":"LDDMM"}. This record does not assert independent reproduction.

### S-RobOT (spline)

Average Error (mm): 5.72. Result ID: 698e6b80ffc3901de09d7355.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Implementation](<https://github.com/uncbiag/robot>)

Recorded evaluation configuration: {"Training regime":"Supervised","Input modality":"Point clouds","Regularization":"Spline"}. This record does not assert independent reproduction.

### CPD (non-rigid)

Average Error (mm): 9.3. Result ID: 698e6b80ffc3901de09d7350.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa603d1e209131ce9ee559.md>)

Recorded evaluation configuration: {"Training regime":"No training","Input modality":"Point clouds","Point cloud density":"20k","Variation":"Non-rigid"}. This record does not assert independent reproduction.

### RobOT (raw)

Average Error (mm): 9.41. Result ID: 698e6b80ffc3901de09d7352.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Implementation](<https://github.com/uncbiag/robot>)

Recorded evaluation configuration: {"Training regime":"No training","Input modality":"Point clouds","Variation":"Raw"}. This record does not assert independent reproduction.

### RobOT (affine)

Average Error (mm): 10.45. Result ID: 698e6b80ffc3901de09d7351.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

[Implementation](<https://github.com/uncbiag/robot>)

Recorded evaluation configuration: {"Training regime":"No training","Input modality":"Point clouds","Variation":"Affine"}. This record does not assert independent reproduction.

### ICP (affine)

Average Error (mm): 15.05. Result ID: 698e6b80ffc3901de09d734f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

Recorded evaluation configuration: {"Training regime":"No training","Input modality":"Point clouds","Variation":"Affine"}. This record does not assert independent reproduction.

### Input data

Average Error (mm): 23.3. Result ID: 698e6b80ffc3901de09d734e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b2f1e209131cea2cad2.md>)

Recorded evaluation configuration: {"Training regime":"No training","Input modality":"Point clouds"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
