Table 4.
Performance comparison (%) with previous methods on FM_RAF-DB and SG_RAF-DB.
| Methods | FM_RAF-DB | SG_RAF-DB |
|---|---|---|
| VGG-16 (Simonyan and Zisserman, 2014) | 73.86 | 75.81 |
| ResNet-50 (He et al., 2015) | 74.32 | 75.88 |
| MobileNetv2 (Sandler et al., 2018) | 73.14 | 75.46 |
| SCN (Wang et al., 2020) | 76.43 | 77.64 |
| MA-Net (Zhao et al., 2021) | 77.64 | 78.78 |
| FFRA-Net(Ours) | 77.87 | 79.50 |
The bold values are outcomes from model runs described in this paper.