3D Point Cloud Classification on ModelNet40 (test) - Adversarial Robustness
0Attack Success RateLPC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LPCAttack Method=FGSM [11], Backbone=VGG162022.03 | 0 | — | |
| LPCAttack Method=JGBA [31], Backbone=VGG162022.03 | 0 | — | |
| LPCAttack Method=FGSM [11], Backbone=ResNet502022.03 | 0 | — | |
| LPCAttack Method=JGBA [31], Backbone=ResNet502022.03 | 0 | — | |
| LPCAttack Method=FGSM [11], Backbone=EfficientNet2022.03 | 0 | — | |
| LPCAttack Method=JGBA [31], Backbone=EfficientNet2022.03 | 0 | — | |
| PointNetAttack Method=Perturbation [62], Defense Method=DUP-Net [75]2022.03 | 0.095 | — | |
| PointNetAttack Method=Add [62], Defense Method=DUP-Net [75]2022.03 | 0.095 | — | |
| PointNetAttack Method=Perturbation [62], Defense Method=IF-Defense [60]2022.03 | 0.095 | — | |
| PointNetAttack Method=Add [62], Defense Method=IF-Defense [60]2022.03 | 0.095 | — | |
| LPCAttack Method=Add [62], Backbone=ResNet502022.03 | 0.19 | — | |
| LPCAttack Method=Add [62], Backbone=EfficientNet2022.03 | 0.19 | — | |
| LPCAttack Method=Perturbation [62], Backbone=ResNet502022.03 | 0.37 | — | |
| LPCAttack Method=Perturbation [62], Backbone=EfficientNet2022.03 | 0.38 | — | |
| LPCAttack Method=Perturbation [62], Backbone=VGG162022.03 | 0.56 | — | |
| LPCAttack Method=Add [62], Backbone=VGG162022.03 | 0.56 | — | |
| LPCAttack Method=Object [62], Backbone=EfficientNet2022.03 | 0.75 | — | |
| LPCAttack Type=LG-GAN [74], Backbone=ResNet2022.03 | 0.8 | 76.7 | |
| LPCAttack Method=Object [62], Backbone=VGG162022.03 | 0.93 | — | |
| LPCAttack Method=Cluster [62], Backbone=ResNet502022.03 | 0.93 | — | |
| PointNetAttack Method=Object [62], Defense Method=DUP-Net [75]2022.03 | 1.02 | — | |
| PointNetAttack Method=Object [62], Defense Method=IF-Defense [60]2022.03 | 1.11 | — | |
| LPCAttack Method=Cluster [62], Backbone=VGG162022.03 | 1.11 | — | |
| LPCAttack Method=Object [62], Backbone=ResNet502022.03 | 1.11 | — | |
| LPCAttack Method=Cluster [62], Backbone=EfficientNet2022.03 | 1.21 | — | |
| RPLAttack Type=LG-GAN [74]2022.03 | 2.6 | 56.8 | |
| PointNetAttack Method=Add [62], Defense Method=RPL [12]2022.03 | 3.33 | — | |
| PointNetAttack Method=Perturbation [62], Defense Method=RPL [12]2022.03 | 3.95 | — | |
| PointNetAttack Method=Cluster [62], Defense Method=IF-Defense [60]2022.03 | 6.3 | — | |
| PointNetAttack Method=Cluster [62], Defense Method=DUP-Net [75]2022.03 | 6.76 | — | |
| IF-DefenseAttack Type=LG-GAN [74]2022.03 | 10.4 | 37.2 | |
| DUP-NetAttack Type=LG-GAN [74]2022.03 | 13.9 | 31.6 | |
| PointNetAttack Method=Cluster [62], Defense Method=RPL [12]2022.03 | 17.04 | — | |
| PointNetAttack Method=FGSM [11], Defense Method=DUP-Net [75]2022.03 | 30.77 | — | |
| PointNetAttack Method=FGSM [11]2022.03 | 48.99 | — | |
| PointNetAttack Method=FGSM [11], Defense Method=IF-Defense [60]2022.03 | 55.78 | — | |
| PointNetAttack Method=Object [62], Defense Method=RPL [12]2022.03 | 74.07 | — | |
| PointNetAttack Method=JGBA [31], Defense Method=IF-Defense [60]2022.03 | 93.85 | — | |
| PointNetAttack Type=LG-GAN [74]2022.03 | 97 | 0.6 | |
| PointNetAttack Method=Cluster [62]2022.03 | 98.34 | — | |
| PointNetAttack Method=Object [62]2022.03 | 98.43 | — | |
| PointNetAttack Method=JGBA [31], Defense Method=DUP-Net [75]2022.03 | 98.73 | — | |
| PointNetAttack Method=Add [62]2022.03 | 99.72 | — | |
| PointNetAttack Method=FGSM [11], Defense Method=RPL [12]2022.03 | 99.95 | — | |
| PointNetAttack Method=JGBA [31]2022.03 | 100 | — | |
| PointNetAttack Method=Perturbation [62]2022.03 | 100 | — | |
| PointNetAttack Method=JGBA [31], Defense Method=RPL [12]2022.03 | 100 | — | |
| APCVictim Model=PointNet2026.04 | — | 86.4 | |
| APCVictim Model=PointNet++2026.04 | — | 87.6 | |
| APCVictim Model=DGCNN2026.04 | — | 86.9 | |
| ATVictim Model=PointNet2026.04 | — | 85.1 | |
| ATVictim Model=PointNet++2026.04 | — | 87 | |
| ATVictim Model=DGCNN2026.04 | — | 87.3 | |
| CausalPCVictim Model=PointNet2026.04 | — | 84.4 | |
| CausalPCVictim Model=DGCNN2026.04 | — | 84.4 | |
| DUP-NetVictim Model=PointNet2026.04 | — | 82.4 | |
| DUP-NetVictim Model=PointNet++2026.04 | — | 88 | |
| DUP-NetVictim Model=DGCNN2026.04 | — | 52.3 | |
| HTVictim Model=PointNet2026.04 | — | 84.3 | |
| HTVictim Model=PointNet++2026.04 | — | 87.1 | |
| HTVictim Model=DGCNN2026.04 | — | 78.4 | |
| IF-DefenseVictim Model=PointNet2026.04 | — | 84.2 | |
| IF-DefenseVictim Model=PointNet++2026.04 | — | 86.9 | |
| IF-DefenseVictim Model=DGCNN2026.04 | — | 87.1 | |
| No DefenseVictim Model=PointNet2026.04 | — | 0 | |
| No DefenseVictim Model=PointNet++2026.04 | — | 84.4 | |
| No DefenseVictim Model=DGCNN2026.04 | — | 6.7 | |
| SORVictim Model=PointNet2026.04 | — | 80.9 | |
| SORVictim Model=PointNet++2026.04 | — | 88 | |
| SORVictim Model=DGCNN2026.04 | — | 87.2 | |
| SRSVictim Model=PointNet2026.04 | — | 81.5 | |
| SRSVictim Model=PointNet++2026.04 | — | 86.9 | |
| SRSVictim Model=DGCNN2026.04 | — | 80.9 |