List your company
ResearchFormerly WizWand

Object Classification on ScanObjectNN OBJ_ONLY v1.0

96.6AccuracyPointGPT-L
78.50483.20287.992.598May 19, 2023Aug 2, 2023Oct 17, 2023Jan 1, 2024Mar 17, 2024Jun 1, 2024Aug 16, 2024
Updated 8mo ago

Evaluation results

MethodLinks
PointGPT-LLearning Paradigm=Self-Supervised Representation Learning, Larger pre-training dataset=true, Post-pre-training stage=true2023.0596.6
PointGPT-BLearning Paradigm=Self-Supervised Representation Learning, Larger pre-training dataset=true, Post-pre-training stage=true2023.0595.2
PCP-MAE#P (M)=22.1, Input=2k Points2024.0894.32
ReCon#P (M)=43.6, Input=2k Points2024.0893.63
ReConLearning Paradigm=Cross-modal Information and Teacher Models2023.0593.6
Point-FEMAE#P (M)=27.4, Input=2k Points2024.0893.29
MVTNLearning Paradigm=Supervised Learning Only2023.0592.3
MVTN#P (M)=11.2, Input=20 Images2024.0892.3
ACT#P (M)=22.1, Input=2k Points2024.0891.91
ACTLearning Paradigm=Cross-modal Information and Teacher Models2023.0591.9
I2P-MAE#P (M)=15.3, Input=2k Points2024.0891.57
PointGPT-SLearning Paradigm=Self-Supervised Representation Learning, Larger pre-training dataset=false, Post-pre-training stage=false2023.0590
PointGPT#P (M)=19.5, Input=2k Points2024.0890
TAP#P (M)=22.1, Input=2k Points2024.0889.5
Joint-MAEInput=2k Points2024.0888.86
Point-M2AE#P (M)=15.3, Input=2k Points2024.0888.81
Point-M2AELearning Paradigm=Self-Supervised Representation Learning2023.0588.8
Point-MAE#P (M)=22.1, Input=2k Points2024.0888.29
Point-MAELearning Paradigm=Self-Supervised Representation Learning2023.0588.2
Point-BERT#P (M)=22.1, Input=1k Points2024.0888.12
Point-BERTLearning Paradigm=Self-Supervised Representation Learning2023.0588.1
MaskPointLearning Paradigm=Self-Supervised Representation Learning2023.0588.1
MaskPointInput=2k Points2024.0888.1
DGCNNLearning Paradigm=Supervised Learning Only2023.0586.2
DGCNN#P (M)=1.8, Input=1k Points2024.0886.2
PointCNNLearning Paradigm=Supervised Learning Only2023.0585.5
PointNet++#P (M)=1.5, Input=1k Points2024.0884.3
PointNetLearning Paradigm=Supervised Learning Only2023.0579.2
PointNet#P (M)=3.5, Input=1k Points2024.0879.2