List your company
ResearchFormerly WizWand

Object Classification on ModelNet40 (Overall)

94.4Overall AccuracyRepSurf-U
88.99290.39691.893.204Jul 27, 2023
Updated 8mo ago

Evaluation results

MethodLinks
RepSurf-U#Params (M)=1.5, Pre-training Strategy=Supervised Learning Only2023.0794.4
PointMLP#Params (M)=12.6, Pre-training Strategy=Supervised Learning Only2023.0794.1
ReCon#Params (M)=44.3, Pre-training Strategy=With Pre-trained Image Model, Extra Knowledge=Pre-trained Image Models2023.0794.1
Point-M2AE#Params (M)=15.3, Pre-training Strategy=Transformers with Pre-training2023.0794
PointMLP + TAP#Params (M)=12.6, Pre-training Strategy=Our Proposed TAP Pre-training2023.0794
SimpleViewPre-training Strategy=Supervised Learning Only2023.0793.9
GBNet#Params (M)=8.8, Pre-training Strategy=Supervised Learning Only2023.0793.8
MVTN#Params (M)=11.2, Pre-training Strategy=Supervised Learning Only2023.0793.8
MaskPoint#Params (M)=22.1, Pre-training Strategy=Transformers with Pre-training2023.0793.8
Point-MAE#Params (M)=22.1, Pre-training Strategy=Transformers with Pre-training2023.0793.8
PRA-Net#Params (M)=2.3, Pre-training Strategy=Supervised Learning Only2023.0793.7
PointNeXt#Params (M)=1.4, Pre-training Strategy=Supervised Learning Only2023.0793.7
ACT#Params (M)=22.1, Pre-training Strategy=With Pre-trained Image Model, Extra Knowledge=Pre-trained Image Models2023.0793.7
I2P-MAEPre-training Strategy=With Pre-trained Image Model, Extra Knowledge=Pre-trained Image Models2023.0793.7
PointMLP (reproduce)#Params (M)=12.6, Pre-training Strategy=TAP Pre-training (Baseline), reproducibility=reproduced by this paper2023.0793.7
Point-BERT#Params (M)=22.1, Pre-training Strategy=Transformers with Pre-training2023.0793.2
DRNetPre-training Strategy=Supervised Learning Only2023.0793.1
DGCNN#Params (M)=1.8, Pre-training Strategy=Supervised Learning Only2023.0792.9
PointCNN#Params (M)=0.6, Pre-training Strategy=Supervised Learning Only2023.0792.2
OcCo#Params (M)=22.1, Pre-training Strategy=Transformers with Pre-training2023.0792.1
Transformer#Params (M)=22.1, Pre-training Strategy=Supervised Learning Only2023.0791.4
PointNet++#Params (M)=1.5, Pre-training Strategy=Supervised Learning Only2023.0790.7
PointNet#Params (M)=3.5, Pre-training Strategy=Supervised Learning Only2023.0789.2