Deepfake Video Detection Based on EfficientNet-V2 Network

Table 4

The performance of the FF++ dataset on various networks.

Mainstream detection network	Accuracy (Acc) (%)
Xcept. full image	74.55
Steg. features	73.64
Bayar and Stamm	84.55
Cozzolino et al.	85.45
Rahmouni et al.	85.45
MesoNet	87.27
XceptionNet	96.36
EfficientNet	97.90