Action Unit Detection on FERA17 (val) benchmark leaderboard · SOTA2 Research

Action Unit Detection on FERA17 (val)

93.3AccuracyOurs

Evaluation Results

Method	Links
Ours2017.04		93.3	48.7
Ours2017.04		92.7	55.6
Ours2017.04		88.1	43.1
Challenge winner2017.04		86.2	47.5
Challenge winner2017.04		86	84
Ours2017.04		85.9	85.3
Ours2017.04		85.7	49.6
Ours2017.04		83.4	64.4
Ours2017.04		83.2	76.1
Challenge winner2017.04		82.9	83.6
Ours2017.04		82.3	83.6
Challenge winner2017.04		82.2	44.2
Ours2017.04		82.2	46.6
Challenge winner2017.04		80.8	36.2
Challenge winner2017.04		80.6	35.3
Challenge winner2017.04		79.9	71.2
Challenge winner2017.04		79.7	58
Challenge winner2017.04		78.2	30.4
Ours2017.04		75.5	81.8
Challenge winner2017.04		73.7	77.9
Baseline approach2017.04		67.6	56.4
Challenge winner2017.04		66.7	69.7
Baseline approach2017.04		66	64.7
Ours2017.04		64.8	73.4
Baseline approach2017.04		64.2	72.7
Baseline approach2017.04		63.8	69.2
Baseline approach2017.04		62.2	62.2
Baseline approach2017.04		57	15.4
Baseline approach2017.04		54.9	41.6
Baseline approach2017.04		52	17.2
Baseline approach2017.04		48.5	22.4
Baseline approach2017.04		37.3	20.7
Baseline approach2017.04		30.7	14.6