Semantic Segmentation On S3Dis

評価指標

Mean IoU

Number of params

oAcc

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名	Mean IoU	Number of params	oAcc	Paper Title	Repository
A-CNN	62.9	N/A	87.3	A-CNN: Annularly Convolutional Neural Networks on Point Clouds
DeepViewAgg	74.7	41.2M	90.1	Learning Multi-View Aggregation In the Wild for Large-Scale 3D Semantic Segmentation
RandLA-Net	-	1.2M	87.1	RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds
MinkowskiNet	65.4	37.9M	-	4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks
MuGNet	69.8	N/A	88.5	MuGNet: Multi-Resolution Graph Neural Network for Large-Scale Pointcloud Segmentation
EQ-Net	77.5	N/A	-	A Unified Query-based Paradigm for Point Cloud Understanding
Feature Geometric Net (FG-Net)	70.8	N/A	88.2	FG-Net: Fast Large-Scale LiDAR Point Clouds Understanding Network Leveraging Correlated Feature Mining and Geometric-Aware Modelling
Swin3D-L	79.8	N/A	92.4	Swin3D: A Pretrained Transformer Backbone for 3D Indoor Scene Understanding
JSENet	67.7	N/A	-	JSENet: Joint Semantic Segmentation and Edge Detection Network for 3D Point Clouds
BIM-Net	-	-	-	Fully Automated Scan-to-BIM Via Point Cloud Instance Segmentation
PointCNN	65.4	N/A	88.1	A-CNN: Annularly Convolutional Neural Networks on Point Clouds
PointASNL	68.7	N/A	88.8	PointASNL: Robust Point Clouds Processing using Nonlocal Neural Networks with Adaptive Sampling
KPConv	70.6	14.1M	-	KPConv: Flexible and Deformable Convolution for Point Clouds
PointCNN	65.4	N/A	-	Point Transformer
BAAF-Net	72.2	N/A	88.9	Semantic Segmentation for Real Point Cloud Scenes via Bilateral Augmentation and Adaptive Fusion
PPT + SparseUNet	78.1	N/A	92.2	Towards Large-scale 3D Representation Learning with Multi-dataset Point Prompt Training
JSNet	61.7	N/A	88.7	JSNet: Joint Instance and Semantic Segmentation of 3D Point Clouds
ShellNet	66.8	N/A	-	ShellNet: Efficient Point Cloud Convolutional Neural Networks using Concentric Shells Statistics	-
SPGraph	62.1	N/A	85.5	A-CNN: Annularly Convolutional Neural Networks on Point Clouds
3P-RNN	56.3	N/A	86.9	A-CNN: Annularly Convolutional Neural Networks on Point Clouds

0 of 54 row(s) selected.