An Improved Self-Training Model with Fine-Tuning Teacher/Student Exchange Strategy to Detect Computer-Generated Images

Table 8

Robustness test results (with attack probability 50%).

Model	Accuracy (%)	Precision (%)	Recall (%)
M0	71.93	86.40	66.99
M1	84.40	91.00	80.38
M2	84.85	90.09	81.53
M3	85.12	90.84	81.51