README.md

---
tags:
- 3D Vision
- NeRF
- 3D Gaussian
- Dataset
- Novel View Synthesis
- Text to 3D
- Image to 3D
pretty_name: Dl3DV-Dataset
size_categories:
- n>1T
---

# DL3DV-GS-960P Dataset 
DL3DV-GS-960P dataset contains 6939 samples of `undistorted images`, `camera poses`, and `pre-trained 3DGS`, under 960P resolution.
This dataset is originated from DL3DV, and post processed by [FCGS](https://github.com/YihangChen-ee/FCGS). More information can be found in the Appendix of [FCGS paper](https://arxiv.org/pdf/2410.08017).

# Download
If you have enough space, you can use git to download a dataset from huggingface.

<!-- The downloading scripts are provided as follows. Please click and download the following files first and put them at the same directory:

1. Python script [download_DL3DV-GS-960P.py](https://github.com/YihangChen-ee/FCGS/blob/main/dataset/download_DL3DV-GS-960P.py) to download the dataset.
2. [Training list](https://github.com/YihangChen-ee/FCGS/blob/main/dataset/hash_name_train.txt) containing hash ids for training scenes. -- Split by [FCGS](https://github.com/YihangChen-ee/FCGS)
3. [Testing list](https://github.com/YihangChen-ee/FCGS/blob/main/dataset/hash_name_test.txt) containing hash ids for testing scenes. -- Split by [FCGS](https://github.com/YihangChen-ee/FCGS)
 -->
Or you can use downloading script that provides more flexibility. 

0. `git clone https://github.com/YihangChen-ee/FCGS.git`. Downloading scripts are put under `dataset` folder in this repo. 
1. Python script `dataset/download_DL3DV-GS-960P.py` to download the dataset.
2. `dataset/hash_name_train.txt` containing hash ids for training scenes. -- Split by [FCGS](https://github.com/YihangChen-ee/FCGS)
3. `dataset/hash_name_test.txt` containing hash ids for testing scenes. -- Split by [FCGS](https://github.com/YihangChen-ee/FCGS)


The usage of the download script: 
```Bash

usage: download_DL3DV-GS-960P.py [-h] --odir ODIR --subset {/,1K,2K,3K,4K,5K,6K,7K} --file_type {3DGS,imgs_undist} [--hash HASH] --split {train,test}

optional arguments:
  -h, --help            show this help message and exit
  --odir ODIR           output directory
  --subset {/,1K,2K,3K,4K,5K,6K,7K}
                        The subset of the benchmark to download. A / by default indicates downloading all subsets one by one. 
  --file_type {3DGS,imgs_undist}
                        The file type to download. 
                        1) 3DGS: pre-trained 3DGS. 
                        2) imgs_undist: both undistorted images and camera poses
  --hash                If set hash, this is the hash code of the scene to download
  --split {train,test}   
                        The training or testing split. For the testing split, chkpnt30000.pth is also provided in the 3DGS file_type.
```

Here are some examples:
```Bash
# Make sure you have applied for the access.

# Download pretrained 3DGS, 1K subset, output to ./DL3DV-GS-960P directory, for training
python download_DL3DV-GS-960P.py --odir ./DL3DV-GS-960P --subset 1K --file_type 3DGS --split train

# Download both undistorted images and camera poses, 6K subset, output to ./DL3DV-GS-960P directory, for testing
python download_DL3DV-GS-960P.py --odir ./DL3DV-GS-960P --subset 6K --file_type imgs_undist --split test
```

You can also download a specific scene with its hash. The scene-hash pair visualization can be found [here](https://htmlpreview.github.io/?https://github.com/DL3DV-10K/Dataset/blob/main/visualize/index.html).

```Bash
python download_DL3DV-GS-960P.py --odir ./DL3DV-GS-960P --hash a38bb446a5e6117ca5cc44fb0809a37ac59a8cfb7093be6d0bc5a5b32aee156e --file_type imgs_undist --split train 
```

## BibTeX
If you found DL3DV-GS-960P Dataset useful, please cite:.

```
@inproceedings{fcgs2025,
  title={Fast Feedforward 3D Gaussian Splatting Compression},
  author={Chen, Yihang and Wu, Qianyi and Li, Mengyao and Lin, Weiyao and Harandi, Mehrtash and Cai, Jianfei},
  booktitle={The Thirteenth International Conference on Learning Representations},
  year={2025}
}
```

```
@inproceedings{ling2024dl3dv,
  title={Dl3dv-10k: A large-scale scene dataset for deep learning-based 3d vision},
  author={Ling, Lu and Sheng, Yichen and Tu, Zhi and Zhao, Wentian and Xin, Cheng and Wan, Kun and Yu, Lantao and Guo, Qianyu and Yu, Zixun and Lu, Yawen and others},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={22160--22169},
  year={2024}
}
```