Channel-Wise Spatiotemporal Aggregation Technology for Face Video Forensics

Table 12

AUC (higher is better) comparison on FF++ datasets.

Approach	DF	F2F	FS	NT	Average
Ours	0.997	1.0	0.999	0.988	0.996
CNN-RNN [3]	0.996	0.984	0.994	—	0.991
Face X-ray [28]	0.992	0.991	0.992	0.989	0.991
Camera noise [9]	0.963	0.939	0.978	—	0.960