List your company
ResearchFormerly WizWand

Facial Expression Recognition on RML (test)

76.53AccuracyARPGNet
59.546863.955968.36572.7741Nov 27, 2025
Updated 8mo ago

Evaluation results

MethodLinks
ARPGNetStream Type=Multi-stream, Backbone=default2025.1176.53
C3D (with audio modality)Stream Type=Single-stream, Modality=Audio + Visual2025.1173.9
Pooling + DBNStream Type=Multi-stream2025.1173.7
MulTStream Type=Multi-stream, Backbone=same as ARPGNet, Fusion Method=MulT2025.1173.33
Two-stageStream Type=Single-stream2025.1172.5
HSE-NNStream Type=Single-stream2025.1172.2
TEMMAStream Type=Multi-stream, Backbone=same as ARPGNet, Fusion Method=TEMMA2025.1171.11
PACVTStream Type=Multi-stream2025.1170.1
MDANStream Type=Single-stream2025.1169.3
AM-CNNStream Type=Multi-stream2025.1168.6
Multi-attentionStream Type=Multi-stream, Backbone=same as ARPGNet, Fusion Method=Multi-attention2025.1168.33
C3DStream Type=Single-stream2025.1168.1
APViTStream Type=Single-stream2025.1167.6
LSTM + LALStream Type=Multi-stream2025.1165.7
Former-DFERStream Type=Single-stream2025.1162.5
PoolingStream Type=Single-stream2025.1160.8
LSTM + RLStream Type=Single-stream2025.1160.2