# Facial Landmark Detection on CHARLOTTE 1.0 (Side)

Source: [Facial Landmark Detection on CHARLOTTE 1.0 (Side)](<https://www.sota2.com/research/sota/facial-landmark-detection-on-charlotte-1-0-side>)

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

Data updated: 2026-02-26T05:53:20.985Z

Evaluation of facial landmark localization accuracy on the CHARLOTTE 1.0 (Side) dataset subset, typically measured using Normalized Mean Error (NME W/H) and Failure Rate.

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 | Facial Landmark Detection |
| Dataset (including any stated split) | CHARLOTTE 1.0 (Side) |
| Metric | NME W/H |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 4 of 4 |
| 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

- [NME W/H](<https://www.sota2.com/research/sota/facial-landmark-detection-on-charlotte-1-0-side.md?metric=NME%20W%2FH>): Lower is better; unit: SCALAR.

- [Failure Rate (%)](<https://www.sota2.com/research/sota/facial-landmark-detection-on-charlotte-1-0-side.md?metric=Failure%20Rate%20(%25)>): Lower is better; unit: PERCENT.

## Recorded results

| Method | NME W/H | Result date | Report paper |
| --- | --- | --- | --- |
| GLL-SW Sparse | 0.0551 | Not recorded | 68fa79151e209131cea7c0a7 |
| GLL-SW Sparse | 0.0593 | Not recorded | 68fa79151e209131cea7c0a7 |
| 3FabRec | 0.0675 | Not recorded | 68fa79151e209131cea7c0a7 |
| Mediapipe | 0.2879 | Not recorded | 68fa79151e209131cea7c0a7 |

### GLL-SW Sparse

NME W/H: 0.0551. Result ID: 6995c4813badfaffe6c79dcd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79151e209131cea7c0a7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79151e209131cea7c0a7.md>)

[Implementation](<https://github.com/phflot/tfake>)

Recorded evaluation configuration: {"Training Dataset":"FAKE + T-FAKE","Confidence threshold":"sigma \< 2.2e-4"}. This record does not assert independent reproduction.

### GLL-SW Sparse

NME W/H: 0.0593. Result ID: 6995c4813badfaffe6c79dcc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79151e209131cea7c0a7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79151e209131cea7c0a7.md>)

[Implementation](<https://github.com/phflot/tfake>)

Recorded evaluation configuration: {"Training Dataset":"FAKE + T-FAKE","Confidence threshold":"sigma \< 6e-4"}. This record does not assert independent reproduction.

### 3FabRec

NME W/H: 0.0675. Result ID: 6995c4813badfaffe6c79dca.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79151e209131cea7c0a7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67961e209131cea173c1.md>)

[Implementation](<https://github.com/browatbn2/3FabRec>)

Recorded evaluation configuration: {"Training Dataset":"300W†","Pre-processing":"None"}. This record does not assert independent reproduction.

### Mediapipe

NME W/H: 0.2879. Result ID: 6995c4813badfaffe6c79dcb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79151e209131cea7c0a7.md>)

Recorded evaluation configuration: {"Training Dataset":"undisclosed","Pre-processing":"None"}. This record does not assert independent reproduction.

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