Emotion Recognition on SEED (Emoji Acc, Label Consistency, Class Metrics)
74.15Positive Class AccuracyFMENet + ResNet-18 classifier
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| FMENet + ResNet-18 classifierTraining dataset=EAV, Fine-tuning=None (Zero-shot), Metric Type=Label Consistency, Channels=28 common electrodes, Protocol=3-way2026.07 | 74.15 | — | 61.82 | 60.23 | 51.07 | 61.82 | |
| FMENet + ResNet-18 classifierTraining dataset=EAV, Fine-tuning=None (Zero-shot), Metric Type=Emoji Accuracy, Channels=28 common electrodes, Protocol=3-way2026.07 | 72.31 | 60.09 | — | 58.37 | 49.58 | 60.09 |