3D Part Segmentation On Shapenet Part

Metrics

Class Average IoU

Instance Average IoU

Results

Performance results of various models on this benchmark

Model Name	Class Average IoU	Instance Average IoU	Paper Title	Repository
InterpCNN	84.0	86.3	Interpolated Convolutional Networks for 3D Point Cloud Understanding	-
Point Voxel Transformer	-	86.5	PVT: Point-Voxel Transformer for Point Cloud Learning	-
Point-JEPA	85.8±0.1	-	Point-JEPA: A Joint Embedding Predictive Architecture for Self-Supervised Learning on Point Cloud	-
SSCNN	82.0	84.7	SyncSpecCNN: Synchronized Spectral CNN for 3D Shape Segmentation	-
Point Cloud Transformer	-	86.4	PCT: Point cloud transformer	-
DensePoint	84.2	86.4	DensePoint: Learning Densely Contextual Representation for Efficient Point Cloud Processing	-
CurveNet	-	86.8	Walk in the Cloud: Learning Curves for Point Clouds Shape Analysis	-
3D-JEPA	86.41	84.93	3D-JEPA: A Joint Embedding Predictive Architecture for 3D Self-Supervised Representation Learning	-
KPConv	85.1	86.4	KPConv: Flexible and Deformable Convolution for Point Clouds	-
GeomGCNN	-	89.1	Exploiting Local Geometry for Feature and Graph Construction for Better 3D Point Cloud Processing with Graph Neural Networks	-
PartNet	84.1	-	PartNet: A Recursive Part Decomposition Network for Fine-grained and Hierarchical Shape Segmentation	-
point2vec	84.6	86.3	Point2Vec for Self-Supervised Representation Learning on Point Clouds	-
RS-CNN	-	86.2	Relation-Shape Convolutional Neural Network for Point Cloud Analysis	-
SGPN	-	85.8	SGPN: Similarity Group Proposal Network for 3D Point Cloud Instance Segmentation	-
P2Sequence	-	85.2	Point2Sequence: Learning the Shape Representation of 3D Point Clouds with an Attention-based Sequence to Sequence Network	-
PointNet	-	83.7	PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation	-
PointGPT	84.8	86.6	-	-
PointGrid	82.2	86.4	PointGrid: A Deep Network for 3D Shape Understanding
DeltaConv (U-ResNet)	-	86.9	DeltaConv: Anisotropic Operators for Geometric Deep Learning on Point Clouds	-
ConvPoint	83.4	85.8	ConvPoint: Continuous Convolutions for Point Cloud Processing	-

0 of 67 row(s) selected.