Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Facial Landmark Detection research benchmarks – Page 3 · SOTA2 Research

Facial Landmark Detection

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Helen 68 landmarks 300-W protocol

4.6Mean Error (%)

8

Feb 26, 2026

WholeBody-Face (WBF)

5.12NME

7

Feb 26, 2026

300-VW Category 3

HGs + SA + GHCU

7.51RMSE

7

Feb 26, 2026

300-W competition 68-points (test)

46.56AUC (Threshold 0.08)

7

Feb 26, 2026

AFLW-frontal (test)

1.81Average Error

7

Feb 26, 2026

CHARLOTTE Low 1.0

0.102NME (W/H)

6

Feb 26, 2026

224 x 224 (test)

1.4Params (M)

6

Feb 26, 2026

300-VW (category C)

Bulat et al. (Ours)

60.1AUCic@0.08

6

Feb 26, 2026

AFLW Full set 68-pt re-annotated (test)

6.01NME

6

Feb 26, 2026

AFLW Frontal 68-pt re-annotated (test)

0.019NME

6

Feb 26, 2026

57Inference Speed (FPS)

5

Feb 26, 2026

CHARLOTTE 1.0 (Side)

0.055NME W/H

4

Feb 26, 2026

24.68Params (M)

4

Feb 26, 2026

CHARLOTTE 1.0 (Full)

0.082NME W/H

3

Feb 26, 2026

CHARLOTTE 1.0 (Front)

0.106NME W/H

3

Feb 26, 2026

12.29Params (M)

3

Feb 26, 2026

18Latency (ms)

3

Feb 26, 2026

Page 3 of 3

Popular datasets

300-W AFLW WFLW 300W CHARLOTTE

Follow for updates

Join SOTA2 Contact