Comparion of F1-scores of four detection methods on fake images generated by CycleGAN, StyleGAN2, and Lightweight GAN, where all datasets are split with 10% training, 10% validation and 80% test data.
| CycleGAN | StyleGAN2 | LightweightGAN | |
|---|---|---|---|
| Zhao et al. | 86.62% | 69.50% | 69.75% |
| Wang et al. | 99.64% | 37.89% | 11.55% |
| ResNet18 | 98.87% | 98.46% | 98.89% |
| ResNet18-FFT | 99.88% | 96.33% | 96.45% |
| Geo-DefakeHop | 99.67% | 99.47% | 99.80% |
| CycleGAN | StyleGAN2 | LightweightGAN | |
|---|---|---|---|
| Zhao et al. | 86.62% | 69.50% | 69.75% |
| Wang et al. | 99.64% | 37.89% | 11.55% |
| ResNet18 | 98.87% | 98.46% | 98.89% |
| ResNet18-FFT | 96.33% | 96.45% | |
| Geo-DefakeHop | 99.67% |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.