Micro-Expression Recognition on SMIC
0.855UF1μ-BERT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| μ-BERT# Classes=32023.04 | 0.855 | 0.8384 | |
| EPIRProtocol=CDE, Params=7.03, FLOPs=84.032026.04 | 0.8279 | 0.8363 | |
| MAE# Classes=32023.04 | 0.8186 | 0.8082 | |
| HTNetProtocol=CDE, Params=438.51, FLOPs=214.62026.04 | 0.8049 | 0.7905 | |
| EPIREvaluation Protocol=SDE protocol, Params=7.03, FLOPs=84.032026.04 | 0.8025 | 0.8213 | |
| SRMCLEvaluation Protocol=SDE protocol2026.04 | 0.7887 | 0.7898 | |
| DSTAN# Classes=32023.04 | 0.78 | 0.77 | |
| MiMaNet# Classes=32023.04 | 0.778 | 0.786 | |
| GLEFFNProtocol=CDE2026.04 | 0.7714 | 0.7856 | |
| AMAN# Classes=32023.04 | 0.77 | 0.7987 | |
| I3D+MOCOEvaluation Protocol=SDE protocol2026.04 | 0.7492 | 0.7561 | |
| EMRProtocol=CDE2026.04 | 0.7461 | 0.753 | |
| EMRNetYear=2019, Type=Deep-Learning, Evaluation Protocol=LOSO, Extra Datasets=CK+2023.04 | 0.746 | 0.753 | |
| MTMNet# Classes=32023.04 | 0.744 | 0.76 | |
| FRL-DGTYear=2022, Type=Deep-Learning, Evaluation Protocol=LOSO2023.04 | 0.743 | 0.749 | |
| FRL-DGTProtocol=CDE2026.04 | 0.743 | 0.749 | |
| MERSiamC3DYear=2021, Type=Deep-Learning, Evaluation Protocol=LOSO2023.04 | 0.736 | 0.76 | |
| MERSiamC3DProtocol=CDE2026.04 | 0.736 | 0.76 | |
| SLSTTEvaluation Protocol=SDE protocol2026.04 | 0.724 | 0.7371 | |
| TSCNN# Classes=32023.04 | 0.7236 | 0.7274 | |
| FGRL-AUFYear=2021, Type=Deep-Learning, Evaluation Protocol=LOSO, Extra Datasets=CASME II and SAMM annotated AUs2023.04 | 0.719 | 0.722 | |
| ME-PLANYear=2022, Type=Deep-Learning, Evaluation Protocol=LOSO, Extra Datasets=CK+, Oulu-CASIA and DFEW macro-expression pre-training2023.04 | 0.713 | 0.726 | |
| DIKD# Classes=32023.04 | 0.71 | 0.7606 | |
| FeatRefProtocol=CDE2026.04 | 0.7011 | 0.7083 | |
| FeatRefYear=2022, Type=Deep-Learning, Evaluation Protocol=LOSO2023.04 | 0.701 | 0.708 | |
| OFF-ApexNetProtocol=CDE2026.04 | 0.6817 | 0.6695 | |
| MFDANProtocol=CDE2026.04 | 0.6815 | 0.7043 | |
| STSTNetProtocol=CDE2026.04 | 0.6801 | 0.7013 | |
| STSTNetYear=2019, Type=Deep-Learning, Evaluation Protocol=LOSO2023.04 | 0.68 | 0.701 | |
| Dual-InceptionProtocol=CDE2026.04 | 0.6645 | 0.6726 | |
| OFVIG-NetProtocol=CDE2026.04 | 0.6435 | 0.64 | |
| RCN-AYear=2020, Type=Deep-Learning, Evaluation Protocol=LOSO2023.04 | 0.633 | 0.644 | |
| RCN-AProtocol=CDE2026.04 | 0.633 | 0.644 | |
| GEMEYear=2021, Type=Deep-Learning, Evaluation Protocol=LOSO2023.04 | 0.629 | 0.657 | |
| GEMEProtocol=CDE2026.04 | 0.629 | 0.657 | |
| Bi-WOOFEvaluation Protocol=SDE protocol2026.04 | 0.62 | 0.622 | |
| MDMOEvaluation Protocol=SDE protocol2026.04 | 0.5845 | 0.5897 | |
| CapsuleNetYear=2019, Type=Deep-Learning, Evaluation Protocol=LOSO2023.04 | 0.582 | 0.588 | |
| CapsuleNetProtocol=CDE2026.04 | 0.582 | 0.5877 | |
| Bi-WOOFYear=2018, Type=Hand-Crafted, Evaluation Protocol=LOSO2023.04 | 0.573 | 0.583 | |
| Bi-WOOFProtocol=CDE2026.04 | 0.5727 | 0.5829 | |
| LBP-TOPEvaluation Protocol=SDE protocol2026.04 | 0.3421 | 0.4338 | |
| LBP-TOPYear=2014, Type=Hand-Crafted, Evaluation Protocol=LOSO2023.04 | 0.2 | 0.528 | |
| LBP-TOPProtocol=CDE2026.04 | 0.2 | 0.528 |