Skip to main content

Table 5 Facial manipulation localization performance in terms of accuracy on Face2Face and FaceSwap datasets with high video quality

From: Multi-attention-based approach for deepfake face and expression swap detection and localization

Method

F2F (HQ)

FS (HQ)

Multi-task [41]

92.8

54.1

LAE [62]

90.9

63.2

ADD [43]

98.33

67.02

Proposed method

94.43

70.04

  1. Bold values indicate the best performace against the specfic dataset in each column