Channel-Wise Spatiotemporal Aggregation Technology for Face Video Forensics

Table 10

AUC (higher is better) comparison on Celeb-DF and DFDC-P datasets.

Approach	Celeb-DF	DFDC-P
Ours	0.997	0.925
Metric learning [26]	0.992	—
Face X-ray [27]	0.748	—
Fakespotter [28]	0.668	—
Mesonet [29]	—	0.753