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 Expression Recognition research benchmarks – Page 4 · SOTA2 Research

Facial Expression Recognition

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

CAS(ME)² Self-reported labels

74.6F1 Score

7

Jun 30, 2026

CAS(ME)² Annotated labels

75.2F1 Score

7

Jun 30, 2026

Occlusion-RAFDB (test)

85.56Accuracy

7

Feb 26, 2026

72.48Accuracy

7

Feb 26, 2026

FERPlus (Occlusion)

93.22Accuracy

7

Jun 9, 2026

Facial Expression Recognition

1.28Parameters (M)

7

Feb 26, 2026

AffectNet 8 cls (val)

63.77Accuracy

7

Feb 26, 2026

CK+ 7 classes (10-fold cross-validation)

99.69Accuracy

7

Feb 26, 2026

78.35Accuracy

6

Apr 1, 2026

98.5mAP

6

Feb 26, 2026

Aff-wild2 (val)

0.498CCC (V)

6

Feb 26, 2026

FER2013 7-class

66.73Accuracy

5

Jun 2, 2026

90.91Accuracy

5

Feb 26, 2026

54.79Accuracy

5

Feb 26, 2026

78.38Accuracy

5

Feb 26, 2026

78.38Accuracy

5

Feb 26, 2026

Occlusion-FERPlus (test)

84.74Accuracy

4

Feb 26, 2026

97.8mAP

4

Feb 26, 2026

96.2Accuracy

4

Feb 26, 2026

88.37WAR

4

Feb 26, 2026

SVM classifier using extracted shape and appearance features

76.91Accuracy

3

May 19, 2026

62.52WAR

2

Feb 26, 2026

Deep CNN using knowledge distillation of AUs and SVM classifier

81.8Accuracy

1

May 19, 2026

Deep CNN using Optical Flow feature extraction

70.18Accuracy

1

May 19, 2026

92mAP

1

Feb 26, 2026

Page 4 of 5

Popular datasets

AffectNet CK+FERplus RAF-DB FER

Follow for updates

Join SOTA2 Contact