Table 7 Component analysis on the proposed detection branch for the high-quality (HQ), low-quality (LQ) FF++, and DFD datasets. Each component is gradually incorporated and evaluated to compare the ACC (%) results
From: Multi-attention-based approach for deepfake face and expression swap detection and localization
ID | Frequency features | Spatial features | SCAB | FF++ (C-23) | FF++ (C-40) | DFD |
|---|---|---|---|---|---|---|
1 | \(\checkmark\) | Â | Â | 87.62 | 82.23 | 60.22 |
2 | \(\checkmark\) | \(\checkmark\) | Â | 92.50 | 89.33 | 72.91 |
4 | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | 97.33 | 92.50 | 76.49 |