Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/I2P-MAE

I2P-MAE

Reported on 25 benchmarks across 4 tasks · 1 paper · 7 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision25 results

Shape Representation Of 3D Point CloudsonModelNet40 10-way (20-shot)
Overall Accuracy· uses extra data· 2022-12-13
95.5
best: 96.5 (ReCon++)
SOTA
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
Shape Representation Of 3D Point CloudsonModelNet40 5-way (10-shot)
Overall Accuracy· uses extra data· 2022-12-13
97
best: 98 (PointGPT)
SOTA
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ClassificationonModelNet40 10-way (20-shot)
Overall Accuracy· uses extra data· 2022-12-13
95.5
best: 96.5 (ReCon++)
SOTA
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ClassificationonModelNet40 5-way (10-shot)
Overall Accuracy· uses extra data· 2022-12-13
97
best: 98 (PointGPT)
SOTA
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud Linear ClassificationonModelNet40
Overall Accuracy· uses extra data· 2022-12-13
93.4
best: 93.6 (ReCon++)
SOTA
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ReconstructiononModelNet40 10-way (20-shot)
Overall Accuracy· uses extra data· 2022-12-13
95.5
best: 96.5 (ReCon++)
SOTA
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ReconstructiononModelNet40 5-way (10-shot)
Overall Accuracy· uses extra data· 2022-12-13
97
best: 98 (PointGPT)
SOTA
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
Shape Representation Of 3D Point CloudsonModelNet40 10-way (20-shot)
Standard Deviation· uses extra data· 2022-12-13
3
best: 13.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
Shape Representation Of 3D Point CloudsonModelNet40 5-way (10-shot)
Standard Deviation· uses extra data· 2022-12-13
1.8
best: 16 (PointNet++)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
Shape Representation Of 3D Point CloudsonModelNet40 10-way (10-shot)
Overall Accuracy· uses extra data· 2022-12-13
92.6
best: 95 (Point-JEPA)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
Shape Representation Of 3D Point CloudsonModelNet40 10-way (10-shot)
Standard Deviation· uses extra data· 2022-12-13
5
best: 13.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
Shape Representation Of 3D Point CloudsonModelNet40 5-way (20-shot)
Overall Accuracy· uses extra data· 2022-12-13
98.3
best: 99.5 (ReCon++)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
Shape Representation Of 3D Point CloudsonModelNet40 5-way (20-shot)
Standard Deviation· uses extra data· 2022-12-13
1.3
best: 15.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ClassificationonModelNet40 10-way (20-shot)
Standard Deviation· uses extra data· 2022-12-13
3
best: 13.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ClassificationonModelNet40 5-way (10-shot)
Standard Deviation· uses extra data· 2022-12-13
1.8
best: 16 (PointNet++)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ClassificationonModelNet40 10-way (10-shot)
Overall Accuracy· uses extra data· 2022-12-13
92.6
best: 95 (Point-JEPA)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ClassificationonModelNet40 10-way (10-shot)
Standard Deviation· uses extra data· 2022-12-13
5
best: 13.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ClassificationonModelNet40 5-way (20-shot)
Overall Accuracy· uses extra data· 2022-12-13
98.3
best: 99.5 (ReCon++)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ClassificationonModelNet40 5-way (20-shot)
Standard Deviation· uses extra data· 2022-12-13
1.3
best: 15.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ReconstructiononModelNet40 10-way (20-shot)
Standard Deviation· uses extra data· 2022-12-13
3
best: 13.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ReconstructiononModelNet40 5-way (10-shot)
Standard Deviation· uses extra data· 2022-12-13
1.8
best: 16 (PointNet++)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ReconstructiononModelNet40 10-way (10-shot)
Overall Accuracy· uses extra data· 2022-12-13
92.6
best: 95 (Point-JEPA)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ReconstructiononModelNet40 10-way (10-shot)
Standard Deviation· uses extra data· 2022-12-13
5
best: 13.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ReconstructiononModelNet40 5-way (20-shot)
Overall Accuracy· uses extra data· 2022-12-13
98.3
best: 99.5 (ReCon++)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785
3D Point Cloud ReconstructiononModelNet40 5-way (20-shot)
Standard Deviation· uses extra data· 2022-12-13
1.3
best: 15.5 (PointNet)
Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoders arXiv:2212.06785