ModelNet

Introduced by Wu et al. in 3D ShapeNets: A Deep Representation for Volumetric Shapes

The ModelNet40 dataset contains synthetic object point clouds. As the most widely used benchmark for point cloud analysis, ModelNet40 is popular because of its various categories, clean shapes, well-constructed dataset, etc. The original ModelNet40 consists of 12,311 CAD-generated meshes in 40 categories (such as airplane, car, plant, lamp), of which 9,843 are used for training while the rest 2,468 are reserved for testing. The corresponding point cloud data points are uniformly sampled from the mesh surfaces, and then further preprocessed by moving to the origin and scaling into a unit sphere.

Source: Geometric Feedback Network for Point Cloud Classification

Homepage

Benchmarks

Add a new result Link an existing benchmark

Task	Dataset Variant	Best Model
3D Point Cloud Classification	ModelNet40	GeomGCNN
Few-Shot 3D Point Cloud Classification	ModelNet40 10-way (20-shot)	ReCon++
Few-Shot 3D Point Cloud Classification	ModelNet40 10-way (10-shot)	ReCon++
Few-Shot 3D Point Cloud Classification	ModelNet40 5-way (20-shot)	ReCon++
Few-Shot 3D Point Cloud Classification	ModelNet40 5-way (10-shot)	PointGPT
3D Point Cloud Linear Classification	ModelNet40	ReCon++
Zero-Shot Transfer 3D Point Cloud Classification	ModelNet40	Uni3D
3D Object Classification	ModelNet40	Ours
3D Object Recognition	ModelNet40	R2-MLP-36
Training-free 3D Point Cloud Classification	ModelNet40	Point-NN
3D Point Cloud Data Augmentation	ModelNet40	AdaPC S,T,R
Zero-Shot Transfer 3D Point Cloud Classification	ModelNet10	ReCon
3D Object Classification	ModelNet10	ORION
3D Object Retrieval	ModelNet40	MVTN
Few-Shot Point Cloud Classification	ModelNet40	ViewNet
Zero-shot 3D Point Cloud Classification	ModelNet40	JM3D+PointMLP