3D Point Cloud Classification on ModelNet40 (test)
94.5OAPointMLP+HyCoRe
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| PointMLP+HyCoReTraining=regularized, voting=true2022.09 | 94.5 | — | 91.9 | |
| PointMLPGroup Equivariant=false2022.12 | 94.5 | — | — | |
| PointMLPLearning Paradigm=Supervised Learning Only, #Params (M)=12.6, Input Data=1k Points, Voting=true2023.12 | 94.5 | — | — | |
| Point-FEMAELearning Paradigm=Single-Modal Self-Supervised Learning, #Params (M)=27.4, Input Data=1k Points, Voting=true2023.12 | 94.5 | — | — | |
| ReconLearning Paradigm=Cross-Modal Self-Supervised Learning, #Params (M)=44.3, Input Data=1k Points, Voting=true2023.12 | 94.5 | — | — | |
| PointMLP+HyCoReTraining=regularized2022.09 | 94.3 | — | 91.7 | |
| CLIP2PointPoints=1,024, Views=102022.10 | 94.2 | — | — | |
| PointMLPInputs=xyz, #Points=10242023.10 | 94.1 | — | — | |
| PointMLPLearning Paradigm=Supervised Learning Only, #Params (M)=12.6, Input Data=1k Points, Voting=false2023.12 | 94.1 | — | — | |
| I2P-MAELearning Paradigm=Cross-Modal Self-Supervised Learning, #Params (M)=15.3, Input Data=1k Points, Voting=true2023.12 | 94.1 | — | — | |
| ReconLearning Paradigm=Cross-Modal Self-Supervised Learning, #Params (M)=44.3, Input Data=1k Points, Voting=false2023.12 | 94.1 | — | — | |
| P2P-HorNetLearning Paradigm=Supervised Learning Only, #Params (M)=195.8, Input Data=40 Images, Voting=false2023.12 | 94 | — | — | |
| Point-M2AELearning Paradigm=Single-Modal Self-Supervised Learning, #Params (M)=15.3, Input Data=1k Points, Voting=true2023.12 | 94 | — | — | |
| Point-FEMAELearning Paradigm=Single-Modal Self-Supervised Learning, #Params (M)=27.4, Input Data=1k Points, Voting=false2023.12 | 94 | — | — | |
| Joint-MAELearning Paradigm=Cross-Modal Self-Supervised Learning, Input Data=1k Points, Voting=false2023.12 | 94 | — | — | |
| P2P: HorNet-LPoints=4,096, Views=402022.10 | 94 | — | — | |
| SimpleView2022.02 | 93.9 | — | — | |
| PointHRInputs=xyz, #Points=10242023.10 | 93.9 | — | — | |
| SimpleViewLearning Paradigm=Supervised Learning Only, Input Data=6 Images, Voting=false2023.12 | 93.9 | — | — | |
| SFRLearning Paradigm=Supervised Learning Only, Input Data=12 Images, Voting=false2023.12 | 93.9 | — | — | |
| PointNet-CSRNumber of points=1024, Voting=false2025.01 | 93.9 | — | 91.1 | |
| DOLPHIN (Best Idea)Backbone=PointNet, Mode=Maximum improvement2025.01 | 93.9 | 91.1 | — | |
| GDANet#params=0.93 M2020.12 | 93.8 | — | — | |
| CurveNet2022.02 | 93.8 | — | — | |
| CurveNetTraining=supervised2022.09 | 93.8 | — | — | |
| Point-MAEProtocol=FULL, Standard Transformer=false, #Point=1k P2022.12 | 93.8 | — | — | |
| MVTNLearning Paradigm=Supervised Learning Only, #Params (M)=11.2, Input Data=12 Images, Voting=false2023.12 | 93.8 | — | — | |
| MaskPointLearning Paradigm=Single-Modal Self-Supervised Learning, #Params (M)=22.1, Input Data=1k Points, Voting=false2023.12 | 93.8 | — | — | |
| Point-MAELearning Paradigm=Single-Modal Self-Supervised Learning, #Params (M)=22.1, Input Data=1k Points, Voting=true2023.12 | 93.8 | — | — | |
| GPSFormerNumber of points=1024, Voting=false2025.01 | 93.8 | — | 91.8 | |
| GPSFormerBackbone=GPSFormer, Design=Human-designed2025.01 | 93.8 | 91.8 | — | |
| MVTNPoints=2,048, Views=122022.10 | 93.8 | — | — | |
| Transformer + Point-MAEPoints=1,024, Views=-2022.10 | 93.8 | — | — | |
| Point TransformerTraining=supervised2022.09 | 93.7 | — | 90.6 | |
| PointMLPTraining=supervised, voting=true2022.09 | 93.7 | — | 91.4 | |
| DGCNN+HyCoReTraining=regularized2022.09 | 93.7 | — | 91 | |
| PointTransformerProtocol=Supervised Learning Only, Standard Transformer=false, #Point=1k P2022.12 | 93.7 | — | — | |
| ACTProtocol=FULL, Standard Transformer=false, #Point=1k P2022.12 | 93.7 | — | — | |
| PTv1Inputs=xyz, #Points=10242023.10 | 93.7 | — | — | |
| ACTLearning Paradigm=Cross-Modal Self-Supervised Learning, #Params (M)=22.1, Input Data=1k Points, Voting=true2023.12 | 93.7 | — | — | |
| I2P-MAELearning Paradigm=Cross-Modal Self-Supervised Learning, #Params (M)=15.3, Input Data=1k Points, Voting=false2023.12 | 93.7 | — | — | |
| PAConv2022.02 | 93.6 | — | — | |
| PA-DGCTraining=supervised2022.09 | 93.6 | — | — | |
| Inductive Bias-aided Transformer (IBT)Input Features=xyz, Point Count=1k2023.04 | 93.6 | 91 | — | |
| DGCNN (2048 points)graph_structure=dynamic, points=2048, k=402018.01 | 93.5 | 90.7 | — | |
| SO-NetAlignment Protocol=Pre-aligned, Input Points=PN50002018.01 | 93.4 | — | — | |
| EATinput=P, #points=1k2021.05 | 93.4 | — | — | |
| GDANet2022.02 | 93.4 | — | — | |
| PointMLPTraining=supervised2022.09 | 93.4 | — | 91.2 | |
| RECONHierarchical=false, Evaluation protocol=Linear SVM, Voting=false2023.02 | 93.4 | — | — | |
| SO-NetInputs=xyz, #Points=50002023.10 | 93.4 | — | — | |
| Point-M2AELearning Paradigm=Single-Modal Self-Supervised Learning, #Params (M)=15.3, Input Data=1k Points, Voting=false2023.12 | 93.4 | — | — | |
| SimpleViewPoints=1,024, Views=62022.10 | 93.4 | — | — | |
| ARGNPL (#)=8, E=true, Params (M)=4.15, Search (Day)=0.22022.05 | 93.33 | — | — | |
| GS-NetInput=2k points, k=322019.12 | 93.3 | — | — | |
| 3DCTNInput Features=xyz, normal, Point Count=1k2023.04 | 93.3 | 91.2 | — | |
| PointStackInput Features=xyz, Point Count=1k2023.04 | 93.3 | 89.6 | — | |
| DensePointnarrowness k=24, network depth L=11, #params=0.67M, #FLOPs/sample=651M2019.09 | 93.2 | — | — | |
| PointManifold (LLE Module)Module=LLE, Points=20482020.10 | 93.2 | 90.2 | — | |
| Centroid Transformer (Ours)MACs(G)=5.4, Params(M)=4.12021.02 | 93.2 | — | — | |
| PCTinput=P, #points=1k2021.05 | 93.2 | — | — | |
| PosPool*Inference Speed (instances/second)=27.62021.10 | 93.2 | — | — | |
| WOLFMixBackbone=DGCNN, Augmentation=WOLFMix2022.02 | 93.2 | — | — | |
| DGCNN+DCGLRTraining=finetuned2022.09 | 93.2 | — | — | |
| DensePointProtocol=Supervised Learning Only, #Point=1k P2022.12 | 93.2 | — | — | |
| PosPoolProtocol=Supervised Learning Only, #Point=5k P2022.12 | 93.2 | — | — | |
| PCTProtocol=Supervised Learning Only, Standard Transformer=false, #Point=1k P2022.12 | 93.2 | — | — | |
| Point-BERTProtocol=FULL, Standard Transformer=true, #Point=1k P2022.12 | 93.2 | — | — | |
| PointASNLInput Features=xyz, normal, Point Count=1k2023.04 | 93.2 | — | — | |
| PCTInput Features=xyz, Point Count=1k2023.04 | 93.2 | — | — | |
| PointNextInputs=xyz, #Points=10242023.10 | 93.2 | — | — | |
| PosPoolInputs=xyz, #Points=50002023.10 | 93.2 | — | — | |
| PCTInputs=xyz, #Points=10242023.10 | 93.2 | — | — | |
| Point-BERTLearning Paradigm=Single-Modal Self-Supervised Learning, #Params (M)=22.1, Input Data=1k Points, Voting=true2023.12 | 93.2 | — | — | |
| Point-MAELearning Paradigm=Single-Modal Self-Supervised Learning, #Params (M)=22.1, Input Data=1k Points, Voting=false2023.12 | 93.2 | — | — | |
| ACTLearning Paradigm=Cross-Modal Self-Supervised Learning, #Params (M)=22.1, Input Data=1k Points, Voting=false2023.12 | 93.2 | — | — | |
| CluRenderEvaluation Protocol=Linear Evaluation, Pre-training Dataset=ShapeNet, Backbone=Transformer2024.04 | 93.2 | — | — | |
| PointNeXtNumber of points=1024, Voting=false2025.01 | 93.2 | — | 90.8 | |
| P2P: ConvNeXt-LPoints=4,096, Views=402022.10 | 93.2 | — | — | |
| Transformer + Point-BERTPoints=1,024, Views=-2022.10 | 93.2 | — | — | |
| PointManifold (MP Module)Module=MP, Points=20482020.10 | 93.1 | 90.1 | — | |
| DRNetinput type=coords, #points=1k2020.05 | 93.1 | — | — | |
| SepNet-W15MACs(G)=22.4, Params(M)=4.82021.02 | 93.1 | — | — | |
| Grid-GCNInference Speed (instances/second)=172.02021.10 | 93.1 | — | — | |
| DGCNN+STRLTraining=finetuned2022.09 | 93.1 | — | — | |
| DRNetProtocol=Supervised Learning Only, #Point=1k P2022.12 | 93.1 | — | — | |
| CloudTransformersInput Features=xyz, Point Count=1k2023.04 | 93.1 | 90.8 | — | |
| P2P: ResNet-101Points=4,096, Views=402022.10 | 93.1 | — | — | |
| Point2NodePoints=10242020.10 | 93 | — | — | |
| PointManifold (LLE Module)Module=LLE, Points=10242020.10 | 93 | 90 | — | |
| PointManifold (MP Module)Module=MP, Points=10242020.10 | 93 | 90.1 | — | |
| RSMixBackbone=DGCNN, Augmentation=RSMix2022.02 | 93 | — | — | |
| PointNet++Backbone=PointNet++, Augmentation=Standard2022.02 | 93 | — | — | |
| PointNet++2022.02 | 93 | — | — | |
| PCT2022.02 | 93 | — | — | |
| RPC2022.02 | 93 | — | — | |
| SGASL (#)=9, E=true, Params (M)=8.87, Search (Day)=0.192022.05 | 92.93 | — | — | |
| DGCNNgraph_structure=dynamic, points=10242018.01 | 92.9 | 90.2 | — | |
| GS-NetInput=1k points2019.12 | 92.9 | — | — | |
| KPConvPoints=10242020.10 | 92.9 | — | — |