ModelNet
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
ModelNet40 point perturbation attack
86.99Accuracy
14
ModelNet40 kNN attack
85.62Accuracy
14
ModelNet40 Drop 100 attack
91.17Accuracy
14
ModelNet40 Drop 200 attack
89.02Accuracy
14
ModelNet40 (official)
93.2Accuracy
13
ModelNet (test)
93.9Accuracy
13
ModelNet40 (test)
63.04Instance Recall (IR)
13
ModelNet-O η=2.5%
76.7OA
13
ModelNet-O η=0.5%
78.7OA
13
ModelNet-O Clean
77.6mAcc
13
ModelNet40 (test)
88.4Accuracy
13
ModelNet40 clean
93.4Overall Accuracy
12
ModelNet target domain (eval)
3.81Chamfer Distance (CD)
12
ModelNet40 Hard
55Top-1 Acc
12
ModelNet40 Medium
55.5Top-1 Acc
12
ModelNet40-LS (test)
62.76Acc (5-way 1-shot)
12
ModelNet40 Clean v1.0
91.3Accuracy
12
ModelNet40 Attack v1.0
91.5Accuracy
12
ModelNet40
0.149Error
12
ModelNet40
99.53AUROC
11
ModelNet40 (test)
3,980W2 Distance
11
ModelNet40 Bookshelf
69.73Number of Nodes
11
ModelNet40 Piano
75.35Node Count
11
ModelNet40 Gaussian Noise twice-sampled (test)
1.305RMSE (R)
11
ModelNet 8k points 40 (test)
0.945Accuracy
11