3D Object Classification on ModelNet10 (test)
99.6Mean Class AccuracyR2-MLP-36
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| R2-MLP-36Views=202022.11 | 99.6 | — | — | — | |
| MVT-smallViews=202022.11 | 99.3 | — | — | — | |
| CAR-NetViews=202022.11 | 99 | — | — | — | |
| RotationNetViews=202022.11 | 98.5 | — | — | — | |
| R2-MLP-36Views=122022.11 | 97.4 | — | — | — | |
| VRN ensemble2017.11 | 97.14 | — | — | — | |
| PointASNLInput Data=pnt, nor, Number of Points=1k2020.03 | 95.9 | — | — | — | |
| R2-MLP-36Views=62022.11 | 95.9 | — | — | — | |
| CAR-NetViews=122022.11 | 95.8 | — | — | — | |
| SO-NetInput Data=pnt, nor, Number of Points=5k2020.03 | 95.7 | — | — | — | |
| PointASNLInput Data=pnt, Number of Points=1k2020.03 | 95.7 | — | — | — | |
| A-CNNInput Data=pnt, nor, Number of Points=1k2020.03 | 95.5 | — | — | — | |
| Latent-GANEvaluation Protocol=Linear SVM on unsupervised representations2017.12 | 95.3 | — | — | — | |
| Point2SequenceInput Data=pnt, Number of Points=1k2020.03 | 95.3 | — | — | — | |
| Relation NetworkViews=122022.11 | 95.3 | — | — | — | |
| MVT-smallViews=122022.11 | 95.3 | — | — | — | |
| 3DmFV-NetNumber of points=1024 or 20482017.11 | 95.2 | — | — | — | |
| PCNNInput Data=pnt, Number of Points=1k2020.03 | 94.9 | — | — | — | |
| SeqViews2SeqLabelsViews=122022.11 | 94.8 | — | — | — | |
| 3D2SeqViewsViews=122022.11 | 94.7 | — | — | — | |
| FoldingNetEvaluation Protocol=Linear SVM on unsupervised representations2017.12 | 94.4 | — | — | — | |
| Ψ-CNN2019.02 | 94.4 | 94.6 | — | — | |
| 3DmFV+VoxNet2017.11 | 94.3 | — | — | — | |
| SO-NetInput Data=pnt, nor, Number of Points=2k2020.03 | 94.1 | — | — | — | |
| Kd-networkNumber of points=327682017.11 | 94 | — | — | — | |
| Kd-NetInput Data=pnt, Number of Points=32k2020.03 | 94 | — | — | — | |
| RotationNetViews=122022.11 | 94 | — | — | — | |
| ORION# Conv=4, # param=4M2016.04 | 93.9 | — | — | — | |
| SO-Net2019.02 | 93.9 | 94.1 | — | — | |
| ORION2017.04 | 93.8 | — | — | — | |
| ORION# Conv=2, # param=910K2016.04 | 93.8 | — | — | — | |
| VRN# Conv=43, # param=18M, setup=single-network results2016.04 | 93.6 | — | — | — | |
| VRN2017.11 | 93.6 | — | — | — | |
| Kd-networkNumber of points=10242017.11 | 93.3 | — | — | — | |
| FusionNet# param=118M2016.04 | 93.1 | — | — | — | |
| FusionNet2017.11 | 93.1 | — | — | — | |
| Kd-Net2019.02 | 92.8 | 93.3 | — | — | |
| VoxNet2017.04 | 92 | — | — | — | |
| VoxNet (baseline)# Conv=2, # param=890K2016.04 | 92 | — | — | — | |
| VoxNet2017.11 | 92 | — | — | — | |
| 3D-GANEvaluation Protocol=Linear SVM on unsupervised representations2017.12 | 91 | — | — | — | |
| ECCInput Data=graphs, Number of Points=1k2020.03 | 90.8 | — | — | — | |
| OctNet2019.02 | 90.1 | 90.9 | — | — | |
| ECCNumber of votes=122017.04 | 90 | 90.8 | — | — | |
| ECC2019.02 | 90 | 90.8 | — | — | |
| ECCNumber of votes=12017.04 | 89.3 | 90 | — | — | |
| DeepPano# Conv=42016.04 | 85.5 | — | — | — | |
| 3DShapeNets2017.04 | 83.5 | — | — | — | |
| ShapeNet# Conv=32016.04 | 83.5 | — | — | — | |
| 3DShapeNets2017.11 | 83.5 | — | — | — | |
| VConv-DAEEvaluation Protocol=Linear SVM on unsupervised representations2017.12 | 80.5 | — | — | — | |
| LFDEvaluation Protocol=Linear SVM on unsupervised representations2017.12 | 79.9 | — | — | — | |
| SPHEvaluation Protocol=Linear SVM on unsupervised representations2017.12 | 79.8 | — | — | — | |
| RECONBackbone=Transformer, Ensemb.=true, Zero-shot=true2023.02 | 75.6 | — | — | — | |
| RECONBackbone=Transformer, Ensemb.=false, Zero-shot=true2023.02 | 74.2 | — | — | — | |
| CLIP2PointBackbone=Transformer, Ensemb.=true, Zero-shot=true2023.02 | 66.6 | — | — | — | |
| PointCLIPBackbone=ResNet-50, Ensemb.=false, Zero-shot=true2023.02 | 30.2 | — | — | — | |
| CFABase Model Mapping=CFA (Fused Model)2026.03 | — | — | — | 88.88 | |
| Cheraghian2D Pre-train=false, 3D Pre-train=true, Zero-shot=true2022.11 | — | — | — | 68.5 | |
| CLIP2Point2D Pre-train=true, 3D Pre-train=false, Zero-shot=true2022.11 | — | — | — | 66.63 | |
| DGCNNBase Model Mapping=A2026.03 | — | — | — | 80.4 | |
| KCNet2019.02 | — | 94.4 | — | — | |
| PointCLIP2D Pre-train=true, 3D Pre-train=false, Zero-shot=true2022.11 | — | — | — | 30.23 | |
| PointCLIP V22D Pre-train=true, 3D Pre-train=false, Zero-shot=true2022.11 | — | — | — | 73.13 | |
| PointCNNBase Model Mapping=B2026.03 | — | — | — | 72.69 | |
| PointNet++variant=SSG, Implementation=PyTorch Geometric2020.02 | — | — | 6.02 | — | |
| PointNet++ + MoExvariant=SSG, p=0.5, lambda=0.9, Normalization=InstanceNorm2020.02 | — | — | 5.25 | — | |
| PointNet2Base Model Mapping=C2026.03 | — | — | — | 84.69 | |
| PointTransformerBase Model Mapping=D2026.03 | — | — | — | 75.33 | |
| SplineCNNBase Model Mapping=E2026.03 | — | — | — | 71.15 |