Table 5.
Comparison between CMC-HF and other state-of-the-art methods for the IEMOCAP dataset.
| Model | Happiness | Sadness | Anger | Neutral | ||||
|---|---|---|---|---|---|---|---|---|
| Acc | F1 | Acc | F1 | Acc | F1 | Acc | F1 | |
| MFN | 86.5 | 84.0 | 83.5 | 82.1 | 85.0 | 83.7 | 69.6 | 69.2 |
| MCTN | 80.5 | 77.5 | 72.0 | 71.4 | 64.9 | 65.6 | 49.4 | 49.3 |
| RAVEN | 77.0 | 76.8 | 67.6 | 65.6 | 65.0 | 64.1 | 62.0 | 59.5 |
| HFusion | 74.3 | 81.4 | 75.6 | 77.0 | 79.6 | 77.6 | 78.4 | 71.2 |
| MulT | 84.8 | 81.9 | 77.7 | 74.4 | 73.9 | 70.2 | 62.5 | 59.7 |
| SSE-FT | 86.5 | 85.7 | 86.7 | 86.2 | 89.4 | 89.0 | 76.0 | 75.9 |
| CMC-HF (ours) | 88.7 | 86.2 | 89.6 | 89.1 | 91.2 | 90.8 | 79.0 | 77.3 |
The values in bold represent the best values of the attributes in this column. The larger the value of Acc and F1, the better the corresponding model.