Visual Emotion Understanding on EmoSet large-scale (full)
85.3AccuracyUniEmo
Evaluation Results
| Method | Links | |
|---|---|---|
| UniEmoBackbone=ViT-L/14, Paradigm=Supervised Emotion Recognition2025.07 | 85.3 | |
| UniEmoBackbone=ViT-B/16, Paradigm=Supervised Emotion Recognition2025.07 | 84.54 | |
| UniEmoBackbone=ViT-B/32, Paradigm=Supervised Emotion Recognition2025.07 | 83.52 | |
| EmoVITBackbone=ViT-L/14, Paradigm=Visual Instruction Tuning2025.07 | 83.36 | |
| MVPBackbone=ViT-B/32, Paradigm=Supervised Emotion Recognition2025.07 | 81.92 | |
| CoCoOpBackbone=ViT-B/32, Paradigm=Supervised Emotion Recognition2025.07 | 80.31 | |
| SimEmotionBackbone=ViT-B/32, Paradigm=Supervised Emotion Recognition2025.07 | 79.06 | |
| PT-DPCBackbone=ViT-B/32, Paradigm=Supervised Emotion Recognition2025.07 | 77.13 | |
| CoOpBackbone=ViT-B/32, Paradigm=Supervised Emotion Recognition2025.07 | 76.19 | |
| MDANBackbone=ResNet-101, Paradigm=Supervised Emotion Recognition2025.07 | 75.75 | |
| BLIP2Backbone=ViT-L/14, Paradigm=Visual Instruction Tuning2025.07 | 46.79 | |
| LLaVABackbone=ViT-L/14, Paradigm=Visual Instruction Tuning2025.07 | 44.03 | |
| InstructBLIPBackbone=ViT-L/14, Paradigm=Visual Instruction Tuning2025.07 | 42.2 | |
| FlamingoBackbone=ViT-L/14, Paradigm=Visual Instruction Tuning2025.07 | 29.59 |