Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/vision-cair/3dcompat-v2
3DCoMPaT++: An improved large-scale 3D vision dataset for compositional recognition
https://github.com/vision-cair/3dcompat-v2
3d compositional-learning computer-vision deep-learning multimodal-deep-learning
Last synced: about 5 hours ago
JSON representation
3DCoMPaT++: An improved large-scale 3D vision dataset for compositional recognition
- Host: GitHub
- URL: https://github.com/vision-cair/3dcompat-v2
- Owner: Vision-CAIR
- License: bsd-3-clause
- Created: 2023-04-14T03:37:52.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-07-09T17:02:27.000Z (4 months ago)
- Last Synced: 2024-11-07T12:46:06.418Z (7 days ago)
- Topics: 3d, compositional-learning, computer-vision, deep-learning, multimodal-deep-learning
- Language: Python
- Homepage:
- Size: 133 MB
- Stars: 76
- Watchers: 7
- Forks: 7
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
3DCoMPaT++: An improved Large-scale 3D Vision Dataset for Compositional Recognition
[](https://arxiv.org/abs/2310.18511)
[](https://colab.research.google.com/drive/1OpgYL_cxekAqZF8B8zuQZkPQxUIxzV0K?usp=sharing)
[](https://3dcompat-dataset.org/doc/)
[](https://3dcompat-dataset.org/doc/dl-dataset.html)
[](https://3dcompat-dataset.org/)
[](https://3dcompat-dataset.org/workshop/)
[](https://eval.ai/web/challenges/challenge-page/2031)
## 📰 News
- **19/08/2023**: As our CVPR23 challenge has finished (congratulations to [Cattalyya Nuengsikapian](https://3dcompat-dataset.org/workshop/#main-section)!), our test set has now been made public. Dataloaders have been updated in consequence: using the "`EvalLoader`" classes is not necessary anymore 😊
- **18/06/2023**: The 3DCoMPaT++ CVPR23 challenge has been concluded. We would like to congratulate [Cattalyya Nuengsikapian](https://3dcompat-dataset.org/workshop/#main-section), winner of both **coarse** and **fine-grained** tracks for her excellent performance in our challenge 🎉
## Summary
- [Introduction](#📚-introduction)
- [Getting started](#🚀-getting-started)
- [Baselines](#📊-baselines)
- [Challenge](#🏆-challenge)
- [Acknowledgments](#🙏-acknowledgments)
- [Citation](#citation)
## 📚 Introduction
3DCoMPaT++ is a multimodal 2D/3D dataset of 16 million rendered views of more than 10 million stylized 3D shapes carefully annotated at **part-instance** level, alongside matching **RGB pointclouds**, **3D textured meshes**, **depth maps** and **segmentation masks**. This work builds upon [3DCoMPaT](https://3dcompat-dataset.org/), the first version of this dataset.
**We plan to further extend the dataset: stay tuned!** 🔥
## 🔍 Browser
To explore our dataset, please check out our integrated web browser:
For more information about the shape browser, please check out [our dedicated Wiki page](https://3dcompat-dataset.org/doc/browser.html).
## 🚀 Getting started
To get started straight away, here is a Jupyter notebook (no downloads required, just **run and play**!):
[](https://colab.research.google.com/drive/1OpgYL_cxekAqZF8B8zuQZkPQxUIxzV0K?usp=sharing)
For a deeper dive into our dataset, please check our online documentation:
[](https://3dcompat-dataset.org/doc/)
## 📊 Baselines
We provide baseline models for 2D and 3D tasks, following the structure below:
- **2D Experiments**
- [2D Shape Classifier](./models/2D/shape_classifier/): ResNet50
- [2D Part and Material Segmentation](./models/2D/segmentation/): SegFormer
- **3D Experiments**
- [3D Shape classification](./models/3D/): DGCNN - PCT - PointNet++ - PointStack - Curvenet - PointNext - PointMLP
- [3D Part segmentation](./models/3D/): PCT - PointNet++ - PointStack - Curvenet - PointNeXT
## 🏆 Challenge
As a part of the [C3DV CVPR 2023 workshop](https://3dcompat-dataset.org/workshop/), we are organizing a modelling challenge based on 3DCoMPaT++.
To learn more about the challenge, check out this link:
[](https://eval.ai/web/challenges/challenge-page/2031)
## 🙏 Acknowledgments
⚙️ For computer time, this research used the resources of the Supercomputing Laboratory at [King Abdullah University of Science & Technology (KAUST)](https://www.kaust.edu.sa/).
We extend our sincere gratitude to the [KAUST HPC Team](www.hpc.kaust.edu.sa) for their invaluable assistance and support during the course of this research project. Their expertise and dedication continues to play a crucial role in the success of our work.
💾 We also thank the [Amazon Open Data](https://aws.amazon.com/opendata) program for providing us with free storage of our large-scale data on their servers. Their generosity and commitment to making research data widely accessible have greatly facilitated our research efforts.
## Citation
If you use our dataset, please cite the two following references:
```bibtex
@article{slim2023_3dcompatplus,
title={3DCoMPaT++: An improved Large-scale 3D Vision Dataset
for Compositional Recognition},
author={Habib Slim, Xiang Li, Yuchen Li,
Mahmoud Ahmed, Mohamed Ayman, Ujjwal Upadhyay
Ahmed Abdelreheem, Arpit Prajapati,
Suhail Pothigara, Peter Wonka, Mohamed Elhoseiny},
year={2023}
}
```
```bibtex
@article{li2022_3dcompat,
title={3D CoMPaT: Composition of Materials on Parts of 3D Things},
author={Yuchen Li, Ujjwal Upadhyay, Habib Slim,
Ahmed Abdelreheem, Arpit Prajapati,
Suhail Pothigara, Peter Wonka, Mohamed Elhoseiny},
journal = {ECCV},
year={2022}
}
```
This repository is owned and maintained by Habib Slim, Xiang Li, Mahmoud Ahmed and Mohamed Ayman, from the Vision-CAIR group.
## References
1. _[Li et al., 2022]_ - 3DCoMPaT: Composition of Materials on Parts of 3D Things.
2. _[Xie et al., 2021]_ - SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers.
3. _[He et al., 2015]_ - Deep Residual Learning for Image Recognition.