</> Code with Papers by Syntology

Datasets › ModelNet

ModelNet

Introduced by Zhirong Wu et al. in 3D ShapeNets: A Deep Representation for Volumetric Shapes1 Jan 2015 archive 2025-07-28

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

HomepageBenchmarksPapers JSON

Benchmarks archive 2025-07-28

All 15 leaderboards whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

First row (archive order)PaperCode
3D Point Cloud Classification ModelNet40 PointGST Overall Accuracy 95.3 Parameter-Efficient Fine-Tuning in Spectral Domain for... jerryfeng2003/pointgst 111 Compare
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (10-shot) Point-JEPA Overall Accuracy 95.0 Point-JEPA: A Joint Embedding Predictive Architecture... Ayumu-J-S/Point-JEPA 31 Compare
Few-Shot 3D Point Cloud Classification ModelNet40 10-way (20-shot) ReCon++ Overall Accuracy 96.5 ShapeLLM: Universal 3D Object Understanding for Embodied... qizekun/ShapeLLM +2 31 Compare
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (10-shot) PointGPT Overall Accuracy 98.0 PointGPT: Auto-regressively Generative Pre-training from... CGuangyan-BIT/PointGPT 30 Compare
Few-Shot 3D Point Cloud Classification ModelNet40 5-way (20-shot) ReCon++ Overall Accuracy 99.5 ShapeLLM: Universal 3D Object Understanding for Embodied... qizekun/ShapeLLM +2 30 Compare
3D Point Cloud Linear Classification ModelNet40 Point-JEPA Overall Accuracy 93.7±0.2 Point-JEPA: A Joint Embedding Predictive Architecture... Ayumu-J-S/Point-JEPA 20 Compare
Zero-Shot Transfer 3D Point Cloud Classification ModelNet40 Uni3D Accuracy (%) 88.2 Uni3D: Exploring Unified 3D Representation at Scale baaivision/uni3d +1 16 Compare
3D Object Classification ModelNet40 Ours Classification Accuracy 93.6 Exploiting Inductive Bias in Transformer for Point Cloud... jiamang/ibt 7 Compare
Training-free 3D Point Cloud Classification ModelNet40 Point-GN Accuracy (%) 85.3 Point-GN: A Non-Parametric Network Using Gaussian... asalarpour/Point_GN 7 Compare
3D Object Recognition ModelNet40 R2-MLP-36 Accuracy 97.7% R2-MLP: Round-Roll MLP for Multi-View 3D Object Recognition shanshuo/MVT +1 6 Compare
Generative 3D Object Classification ModelNet40 MiniGPT-3D ModelNet40 (Average) 60.86 MiniGPT-3D: Efficiently Aligning 3D Point Clouds with... tangyuan96/minigpt-3d 6 Compare
3D Object Classification ModelNet10 PolyNet Accuracy 94.93 PolyNet: Polynomial Neural Network for 3D Shape... myavartanoo/PolyNet_PyTorch 4 Compare
Zero-Shot Transfer 3D Point Cloud Classification ModelNet10 ReCon Accuracy (%) 75.6 Contrast with Reconstruct: Contrastive 3D Representation... qizekun/ReCon +4 4 Compare
3D Object Retrieval ModelNet40 MVTN Mean AP 92.9 MVTN: Multi-View Transformation Network for 3D Shape Recognition ajhamdi/MVTN +1 1 Compare
Few-Shot Point Cloud Classification ModelNet40 ViewNet Top-1 Accuracy(5-Way-1-Shot) 82.87 ViewNet: A Novel Projection-Based Backbone With View... jiajingchen113322/ViewNet 1 Compare

Papers archive 2025-07-28

30 shown of 140 papers with a leaderboard row on this dataset's benchmarks, newest first. The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset; the archive's count for this dataset is 1,406. The Syntology column is from Syntology's graph (read 2026-09-24), stated per sample; it is not part of any archive number.

DateSamples run Syntology
Asymmetric Dual Self-Distillation for 3D Self-Supervised Representation Learning 1 1 26 Jun 2025 not harvested
AdaCrossNet: Adaptive Dynamic Loss Weighting for Cross-Modal Contrastive Point Cloud Learning 1 2 2 Jan 2025 not harvested
Rethinking Masked Representation Learning for 3D Point Cloud Understanding 1 6 26 Dec 2024 not harvested
Point-GN: A Non-Parametric Network Using Gaussian Positional Encoding for Point Cloud Classification 1 1 4 Dec 2024 not harvested
Parameter-Efficient Fine-Tuning in Spectral Domain for Point Cloud Learning 1 1 10 Oct 2024 ran 12 of 13 samples (1 unverified)
3D-JEPA: A Joint Embedding Predictive Architecture for 3D Self-Supervised Representation Learning 0 5 24 Sep 2024 not harvested
Positional Prompt Tuning for Efficient 3D Representation Learning 1 1 21 Aug 2024 not harvested
PCP-MAE: Learning to Predict Centers for Point Masked Autoencoders 1 5 16 Aug 2024 not harvested
MiniGPT-3D: Efficiently Aligning 3D Point Clouds with Large Language Models using 2D Priors 1 1 2 May 2024 not harvested
Point-JEPA: A Joint Embedding Predictive Architecture for Self-Supervised Learning on Point Cloud 1 7 25 Apr 2024 not harvested
OpenDlign: Open-World Point Cloud Understanding with Depth-Aligned Images 1 4 25 Apr 2024 ran 4 of 4 samples (0 unverified; 4 pointer-only for licence)
Mamba3D: Enhancing Local Features for 3D Point Cloud Analysis via State Space Model 1 1 23 Apr 2024 ran 6 of 7 samples (1 unverified; 7 pointer-only for licence)
ShapeLLM: Universal 3D Object Understanding for Embodied Interaction 3 9 27 Feb 2024 ran 9 of 17 samples (8 unverified)
ModelNet-O: A Large-Scale Synthetic Dataset for Occlusion-Aware Point Cloud Classification 1 1 16 Jan 2024 not harvested
Towards Compact 3D Representations via Point Feature Enhancement Masked Autoencoders 1 3 17 Dec 2023 not harvested
Sculpting Holistic 3D Representation in Contrastive Language-Image-3D Pre-training 1 1 3 Nov 2023 ran 0 of 3 samples (3 unverified; 3 pointer-only for licence)
Uni3D: Exploring Unified 3D Representation at Scale 2 1 10 Oct 2023 ran 1 of 2 samples (1 unverified)
DualMLP: a two-stream fusion model for 3D point cloud classification 1 1 10 Oct 2023 not harvested
Regress Before Construct: Regress Autoencoder for Point Cloud Self-supervised Learning 1 5 25 Sep 2023 not harvested
Point-Bind & Point-LLM: Aligning Point Cloud with Multi-modality for 3D Understanding, Generation, and Instruction Following 5 1 1 Sep 2023 ran 13 of 20 samples (7 unverified; 7 pointer-only for licence)
PointLLM: Empowering Large Language Models to Understand Point Clouds 3 2 31 Aug 2023 not harvested
Decoupled Local Aggregation for Point Cloud Learning 1 1 31 Aug 2023 not harvested
ViT-Lens: Initiating Omni-Modal Exploration through 3D Insights 1 1 20 Aug 2023 not harvested
ExpPoint-MAE: Better interpretability and performance for self-supervised point cloud transformers 1 1 19 Jun 2023 not harvested
Point-LGMask: Local and Global Contexts Embedding for Point Cloud Pre-training with Multi-Ratio Masking 1 4 8 Jun 2023 not harvested
CrossMoCo: Multi-modal Momentum Contrastive Learning for Point Cloud 1 7 8 Jun 2023 not harvested
PointGPT: Auto-regressively Generative Pre-training from Point Clouds 1 4 19 May 2023 ran 5 of 7 samples (2 unverified)
OpenShape: Scaling Up 3D Shape Representation Towards Open-World Understanding 1 2 18 May 2023 ran 0 of 7 samples (7 unverified)
Exploiting Inductive Bias in Transformer for Point Cloud Classification and Segmentation 1 1 27 Apr 2023 not harvested
Instance-aware Dynamic Prompt Tuning for Pre-trained Point Cloud Models 3 5 14 Apr 2023 ran 9 of 14 samples (5 unverified; 14 pointer-only for licence)

The full list of 140 is in the JSON twin.

Dataset loaders archive 2025-07-28

pytorch-geometric.readthedocs.io/en/latest/modules/datasets.htmlrusty1s/pytorch_geometricpytorch
graphneural.network/datasets/#modelnetdanielegrattarola/spektraltf

2 loaders as listed in the archive; links are outbound and not re-checked here.

Tasks archive 2025-07-28

3D Point Cloud ClassificationFew-Shot 3D Point Cloud ClassificationZero-shot 3D Point Cloud Classification3D Object Classification3D Object Recognition3D Point Cloud Data AugmentationZero-Shot Transfer 3D Point Cloud ClassificationFew-Shot Point Cloud Classification3D Object Retrieval3D Point Cloud Linear ClassificationTraining-free 3D Point Cloud ClassificationGenerative 3D Object Classification3D Parameter-Efficient Fine-Tuning for Classification

License archive 2025-07-28

Unknown

Modalities archive 2025-07-28

3DPoint cloud

Languages archive 2025-07-28

Chinese

Variants archive 2025-07-28

  • ModelNet40
  • ModelNet10
  • ModelNet
  • ModelNet40 5-way (10-shot)
  • ModelNet40 5-way (20-shot)
  • ModelNet40 10-way (10-shot)
  • ModelNet40 10-way (20-shot)

7 variant names, as the archive lists them.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections