Model Inversion Attack on CelebA private FFHQ public (face.evoLve test)
94Top-5 Attack AccuracyKEDMI + LOMMA
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| KEDMI + LOMMATarget Model=face.evoLve, Evaluation Model=face.evoLve, Public Dataset=FFHQ, Private Dataset=CelebA2023.04 | 94 | 13.33 | 47.51 | |
| KEDMI + LOMTarget Model=face.evoLve, Evaluation Model=face.evoLve, Public Dataset=FFHQ, Private Dataset=CelebA2023.04 | 91.33 | 10.67 | 47.3 | |
| GMI + LOMMATarget Model=face.evoLve, Evaluation Model=face.evoLve, Public Dataset=FFHQ, Private Dataset=CelebA2023.04 | 89 | 55.67 | 40.03 | |
| KEDMI + MATarget Model=face.evoLve, Evaluation Model=face.evoLve, Public Dataset=FFHQ, Private Dataset=CelebA2023.04 | 88.67 | 8 | 35.94 | |
| KEDMITarget Model=face.evoLve, Evaluation Model=face.evoLve, Public Dataset=FFHQ, Private Dataset=CelebA2023.04 | 80.67 | — | 38.09 | |
| GMI + LOMTarget Model=face.evoLve, Evaluation Model=face.evoLve, Public Dataset=FFHQ, Private Dataset=CelebA2023.04 | 74.67 | 41.33 | 44.01 | |
| GMI + MATarget Model=face.evoLve, Evaluation Model=face.evoLve, Public Dataset=FFHQ, Private Dataset=CelebA2023.04 | 72 | 38.67 | 35.58 | |
| GMITarget Model=face.evoLve, Evaluation Model=face.evoLve, Public Dataset=FFHQ, Private Dataset=CelebA2023.04 | 33.33 | — | 52.84 |