Facial Action Unit Classification on EmotioNet (val) benchmark leaderboard · SOTA2 Research

Facial Action Unit Classification on EmotioNet (val)

0.818AUC (AU 1)VGG-Face descriptor

Evaluation Results

Method	Links
VGG-Face descriptor2018.08		0.818	0.83	0.835	0.818	0.92	0.909	0.957	0.806	0.852	0.865	0.73	0.849
VGG-112018.08		0.747	0.772	0.858	0.837	0.938	0.897	0.975	0.783	0.869	0.964	0.815	0.86
FAb-Net2018.08		0.741	0.723	0.758	0.688	0.907	0.818	0.925	0.737	0.772	0.836	0.736	0.786
FAb-Net2018.08		0.734	0.718	0.753	0.678	0.904	0.788	0.919	0.724	0.745	0.837	0.733	0.776
FAb-Net2018.08		0.72	0.689	0.732	0.694	0.882	0.786	0.895	0.71	0.759	0.814	0.72	0.764
Gidaris et al.2018.08		0.686	0.64	0.728	0.7	0.839	0.781	0.838	0.684	0.726	0.731	0.672	0.729
Autoencoder2018.08		0.672	0.605	0.701	0.651	0.796	0.704	0.801	0.683	0.665	0.705	0.641	0.693
SplitBrain2018.08		0.655	0.598	0.667	0.608	0.718	0.658	0.733	0.645	0.574	0.681	0.611	0.65