Table 3

Detection accuracy for DeepFor and VFHQ datasets for our method and the compared methods.

No. train videos (c)DeepForVFHQ
 F -SVMXceptDeFakeHopResNetLSTMF-SVMXceptDeFakeHopResNetLSTM
267.550.061.365.961.750.060.658.4
474.850.062.870.675.550.063.877.5
891.450.062.085.984.850.069.993.3
1695.553.072.891.790.173.672.690.8
3296.556.079.897.192.391.077.888.6
6496.883.085.699.093.594.980.988.3
12896.892.388.598.094.495.579.893.2
25697.098.589.799.996.294.482.794.3

or Create an Account

Close Modal
Close Modal