File size: 6,872 Bytes
c3378d3 8ed7820 c3378d3 1a69e96 c3378d3 8ed7820 c3378d3 8ed7820 1a69e96 c3378d3 6ad5556 c3378d3 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 6ad5556 8ed7820 c3378d3 8ed7820 6ad5556 c3378d3 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 |
---
license: apache-2.0
language:
- en
pipeline_tag: image-to-3d
modalities:
- image
- point clouds
- mesh
arxiv: 2411.14384
---
# [ICCV 2025] DiffusionGS: Baking Gaussian Splatting into Diffusion Denoiser for Fast and Scalable Single-stage Image-to-3D Generation and Reconstruction
## Model Description
These three models are trained for image-to-3D generation on object- and scene-level with the spatial resolution of 256x256 and 512x512. For object-level 3D generation,
mesh exportation is also supported. Here are some generated examples:
路 (a) Object-level Generation
<p align="center">
<table border="0" cellspacing="0" cellpadding="0" style="border-collapse:collapse;border:0;">
<!-- ===== Row 1 ===== -->
<tr>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/abo.gif" width="210">
</td>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/gso.gif" width="210">
</td>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/real_img.gif" width="210">
</td>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/wild.gif" width="210">
</td>
</tr>
<!-- ===== Row 2 ===== -->
<tr>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/sd_2.gif" width="210">
</td>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/sd_1.gif" width="210">
</td>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/flux_1.gif" width="210">
</td>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/green_man.gif" width="210">
</td>
</tr>
</table>
</p>
路 (b) Mesh Exportation
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/mesh.png" width="880">
路 (c) Scene-level Generation
<p align="center">
<table border="0" cellspacing="0" cellpadding="0" style="border-collapse:collapse;border:0;">
<!-- ===== Row 1 ===== -->
<tr>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/plaza.gif" width="430">
</td>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/town.gif" width="410">
</td>
</tr>
<!-- ===== Row 2 ===== -->
<tr>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/cliff.gif" width="425">
</td>
<td style="border:0;padding:10px;">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/art_gallery.gif" width="415">
</td>
</tr>
</table>
</p>
路 (d) Comparison with Hunyuan3D-v2.5
The first row is the prompt image. The second row is Hunyuan3D-v2.5. The third row is our DiffusionGS.
Our method generates better results while enjoying 7.5x faster inference speed.
<p align="center">
<table border="0" cellspacing="0" cellpadding="0" style="border-collapse:collapse;border:0;">
<!-- ===== Row 1: Prompt Image ===== -->
<tr>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/1.png" width="300">
</td>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/2.jpg" width="300">
</td>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/3.png" width="300">
</td>
</tr>
<tr>
<td colspan="3" align="center" style="padding-top:0px;font-style:italic;font-size:24px;">
Prompt Images at Any Viewpoints
</td>
</tr>
<!-- ===== Row 2: Hunyuan3D ===== -->
<tr>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/hunyuan_1.gif" width="300">
</td>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/hunyuan_2.gif" width="300">
</td>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/hunyuan_3.gif" width="300">
</td>
</tr>
<tr>
<td colspan="3" align="center" style="padding-top:0px;font-style:italic;font-size:24px;">
Tencent Hunyuan3D-v2.5 (Inference Time: 180 seconds)
</td>
</tr>
<!-- ===== Row 3: DiffusionGS ===== -->
<tr>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/ours_1.gif" width="300">
</td>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/ours_2.gif" width="300">
</td>
<td style="border:0;padding:10px;" align="center">
<img src="https://raw.githubusercontent.com/caiyuanhao1998/Open-DiffusionGS/master/img/ours_3.gif" width="300">
</td>
</tr>
<tr>
<td colspan="3" align="center" style="padding-top:0px;font-style:italic;font-size:24px;">
Our DiffusionGS (Inference Time: 24 seconds)
</td>
</tr>
</table>
</p>
## Github Code Link
Please refer to our GitHub repo for more detailed instructions on using our code and models.
https://github.com/caiyuanhao1998/Open-DiffusionGS/
## Project Page Link
For more video and interactive generation results, please refer to our project page:
https://caiyuanhao1998.github.io/project/DiffusionGS/
## Arxiv Paper Link
For more technical details, please refer to our ICCV 2025 paper:
https://arxiv.org/abs/2411.14384
## Citation
If you find our code, data, and models useful, please consider citing our paper:
```sh
@inproceedings{diffusiongs,
title={Baking Gaussian Splatting into Diffusion Denoiser for Fast and Scalable Single-stage Image-to-3D Generation and Reconstruction},
author={Yuanhao Cai and He Zhang and Kai Zhang and Yixun Liang and Mengwei Ren and Fujun Luan and Qing Liu and Soo Ye Kim and Jianming Zhang and Zhifei Zhang and Yuqian Zhou and Yulun Zhang and Xiaokang Yang and Zhe Lin and Alan Yuille},
booktitle={ICCV},
year={2025}
}
``` |