Adversarial Attack on ImageNet 1k (test)
39.01AccuracySCGA
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SCGATarget Model/Defense=R&P2025.06 | 39.01 | 51.63 | 57.86 | 9.08 | |
| SCGATarget Model/Defense=BDR2025.06 | 44.07 | 45.37 | 52.22 | 10.29 | |
| Zhang et al. (2022b)Target Model/Defense=R&P2025.06 | 44.78 | 44.59 | 51.6 | 10.56 | |
| SCGATarget Model/Defense=Adv.ViT2025.06 | 45.33 | 11.95 | 25.31 | 4.57 | |
| Zhang et al. (2022b)Target Model/Defense=Adv.ViT2025.06 | 45.64 | 11.72 | 25.48 | 4.96 | |
| Zhang et al. (2022b)Target Model/Defense=BDR2025.06 | 47.82 | 40.76 | 48.06 | 11.3 | |
| BenignTarget Model/Defense=Adv.ViT2025.06 | 48.82 | — | — | — | |
| SCGATarget Model/Defense=Avg.2025.06 | 51.8 | 26.52 | 35.28 | 9.01 | |
| SCGATarget Model/Defense=Adv.ConvNeXt2025.06 | 53.62 | 10.65 | 19.6 | 3.38 | |
| Zhang et al. (2022b)Target Model/Defense=Adv.ConvNeXt2025.06 | 53.88 | 10.26 | 19.4 | 3.46 | |
| Zhang et al. (2022b)Target Model/Defense=Avg.2025.06 | 54.03 | 23.75 | 32.78 | 9.51 | |
| BenignTarget Model/Defense=Adv.ConvNeXt2025.06 | 58.44 | — | — | — | |
| SCGATarget Model/Defense=JPEG2025.06 | 60.83 | 23.74 | 31.61 | 11.48 | |
| Zhang et al. (2022b)Target Model/Defense=JPEG2025.06 | 63.49 | 20.24 | 28.09 | 11.45 | |
| SCGATarget Model/Defense=Adv.IncV32025.06 | 67.92 | 15.75 | 24.83 | 15.23 | |
| BenignTarget Model/Defense=Avg.2025.06 | 68.26 | — | — | — | |
| Zhang et al. (2022b)Target Model/Defense=Adv.IncV32025.06 | 68.54 | 14.95 | 24.02 | 15.3 | |
| BenignTarget Model/Defense=JPEG2025.06 | 74.68 | — | — | — | |
| BenignTarget Model/Defense=BDR2025.06 | 74.68 | — | — | — | |
| BenignTarget Model/Defense=Adv.IncV32025.06 | 76.33 | — | — | — | |
| BenignTarget Model/Defense=R&P2025.06 | 76.58 | — | — | — |