File size: 10,607 Bytes
091f734
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
---
license: mit
datasets:
- Somayeh-h/Nordland
- OPR-Project/OxfordRobotCar_OpenPlaceRecognition
language:
- en
metrics:
- recall_at_1
- recall_at_5
pipeline_tag: image-feature-extraction
tags:
- place-recognition
- visual-place-recognition
- computer-vision
- transformer
- 3d-vision
library:
- pytorch
- lightning
---

# Model Card for UniPR-3D

UniPR-3D is a universal visual place recognition (VPR) framework that supports both **single-frame** and **sequence-to-sequence** matching. It leverages **3D visual geometry grounded tokens** within a transformer architecture to produce robust, viewpoint-invariant descriptors for long-term place recognition under challenging environmental variations (e.g., seasonal, weather, lighting, and viewpoint changes).

## Model Details

### Model Description

- **Developed by:** Tianchen Deng, Xun Chen, Ziming Li, Hongming Shen, Danwei Wang, Javier Civera, Hesheng Wang  
- **Shared by:** Tianchen Deng  
- **Model type:** Vision Transformer with 3D-aware token aggregation for visual place recognition  
- **Language(s):** English (dataset metadata); model is vision-only  
- **License:** MIT  

### Model Sources

- **Repository:** [repo](https://github.com/dtc111111/UniPR-3D)  
- **Paper:** [UniPR-3D: Towards Universal Visual Place Recognition with 3D Visual Geometry Grounded Transformer](https://arxiv.org/abs/2512.21078) (arXiv:2512.21078, 2025)  
- **Demo:** No demo available

## Uses

### Direct Use

This model can be used **out-of-the-box** to extract compact, discriminative global descriptors from:
- Single RGB images (for frame-to-frame VPR)
- Sequences of images (for sequence-to-sequence VPR)

These descriptors are suitable for large-scale localization, robot navigation, and SLAM systems requiring robustness to appearance changes.

### Downstream Use

- Integration into **visual SLAM** or **long-term autonomous navigation** pipelines  
- Replacement for traditional VPR backbones (e.g., NetVLAD, MixVPR, EigenPlaces)  
- Fine-tuning on domain-specific datasets (e.g., underground, aerial, or underwater environments)

### Out-of-Scope Use

- **Not intended** for real-time inference on low-power embedded devices without optimization (latency ~8.23 ms on RTX 4090)  
- **Not designed** for non-visual modalities (e.g., LiDAR, audio, text)  
- Performance may degrade in **extreme occlusion**, **textureless scenes**, or **indoor environments not seen during training**

## Bias, Risks, and Limitations

- Trained primarily on **urban street-level imagery** (GSV-Cities, Mapillary MSLS), so generalization to rural, indoor, or non-Western cities may be limited  
- Inherits biases from training data (e.g., geographic overrepresentation of North America/Europe)  
- No explicit fairness or demographic considerations (as it is a geometric vision model)

### Recommendations

- Evaluate on target domain before deployment  
- Monitor recall performance on your specific dataset using standard VPR metrics (R@1, R@5)

## How to Get Started with the Model

The exact inference script is provided in the GitHub repo (`eval_lora.py`, `main_ft.py`). Pretrained weights are available on Hugging Face or via the repo release.

## Training Details

### Training Data

- **Single-frame model**: Trained on [GSV-Cities](https://github.com/amaralibey/gsv-cities)  
- **Multi-frame model**: Trained on [Mapillary Street-Level Sequences (MSLS)](https://www.mapillary.com/dataset/places)  
- Both datasets contain millions of geo-tagged urban street-view images across diverse cities, seasons, and conditions.

### Training Procedure

#### Preprocessing
- Images resized to 518Γ—518  
- Sequences sampled with spatial proximity for multi-frame training

#### Training Hyperparameters
- **Backbone**: DINOv2 (ViT-large)  
- **Optimization**: AdamW, learning rate scheduling  
- **Loss**: Multi-similarity loss with pair weighting  
- **Training regime**: Mixed-precision (fp16) on NVIDIA GPUs

#### Speeds, Sizes, Times
- **Inference latency**: Single frame - 8.23 ms per image (RTX 4090)  
- **Descriptor dimension**: 17152 (for UniPR-3D)  
- Training time: Not disclosed (multi-day runs on multi-GPU setup)

## Evaluation

### Testing Data, Factors & Metrics

#### Testing Data
- Single frame evaluation:
  - <a href="https://codalab.lisn.upsaclay.fr/competitions/865">MSLS Challenge</a>, where you upload your predictions to their server for evaluation.
  - Single-frame <a href="https://www.mapillary.com/dataset/places">MSLS</a> Validation set
  - Nordland dataset, <a href="https://data.ciirc.cvut.cz/public/projects/2015netVLAD/Pittsburgh250k/">Pittsburgh</a> dataset and SPED dataset, you may download them from <a href="https://surfdrive.surf.nl/index.php/s/sbZRXzYe3l0v67W">here</a>, aligned with DINOv2 SALAD.
- Multi-frame evaluation:
  - Multi-frame <a href="https://www.mapillary.com/dataset/places">MSLS</a> Validation set
  - Two sequence from <a href="https://robotcar-dataset.robots.ox.ac.uk/datasets/">Oxford RobotCar</a>, you may download them <a href="https://entuedu-my.sharepoint.com/personal/heshan001_e_ntu_edu_sg/_layouts/15/onedrive.aspx?id=%2Fpersonal%2Fheshan001%5Fe%5Fntu%5Fedu%5Fsg%2FDocuments%2Fcasevpr%5Fdatasets%2Foxford%5Frobotcar&viewid=e5dcb0e9%2Db23f%2D44cf%2Da843%2D7837d3064c2e&ga=1">here</a>.
    - 2014-12-16-18-44-24 (winter night) query to 2014-11-18-13-20-12 (fall day) db
    - 2014-11-14-16-34-33 (fall night) query to 2015-11-13-10-28-08 (fall day) db
  - <a href="https://github.com/gmberton/VPR-datasets-downloader/blob/main/download_nordland.py">Nordland (filtered) dataset</a>

#### Factors
- Seasonal variation (summer ↔ winter)
- Day vs. night
- Weather (sunny, rainy, snowy)
- Viewpoint change (lateral shift, orientation)

#### Metrics
- **Recall@K (R@1, R@5, R@10)**: Standard metric for VPR – fraction of queries with correct match in top-K retrieved database images

### Results

#### Summary

Our method achieves significantly higher recall than competing approaches, achieving new state-of-the-art performance on both single and multiple frame benchmarks.
##### Single-frame matching results

<style>
  table, th, td {
    border-collapse: collapse;
    text-align: center;
  }
</style>
<table>
  <tr>
    <th colspan="2"></th>
    <th colspan="2">MSLS Challenge</th>
    <th colspan="2">MSLS Val</th>
    <th colspan="2">NordLand</th>
    <th colspan="2">Pitts250k-test</th>
    <th colspan="2">SPED</th>
  </tr>
  <tr>
    <th>Method</th>
    <th>Latency (ms)</th>
    <th>R@1</th>
    <th>R@5</th>
    <th>R@1</th>
    <th>R@5</th>
    <th>R@1</th>
    <th>R@5</th>
    <th>R@1</th>
    <th>R@5</th>
    <th>R@1</th>
    <th>R@5</th>
  </tr>
  <tr>
    <td>MixVPR</td>
    <td>1.37</td>
    <td>64.0</td>
    <td>75.9</td>
    <td>88.0</td>
    <td>92.7</td>
    <td>58.4</td>
    <td>74.6</td>
    <td>94.6</td>
    <td><u>98.3</u></td>
    <td>85.2</td>
    <td>92.1</td>
  </tr>
  <tr>
    <td>EigenPlaces</td>
    <td>2.65</td>
    <td>67.4</td>
    <td>77.1</td>
    <td>89.3</td>
    <td>93.7</td>
    <td>54.4</td>
    <td>68.8</td>
    <td>94.1</td>
    <td>98.0</td>
    <td>69.9</td>
    <td>82.9</td>
  </tr>
  <tr>
    <td>DINOv2 SALAD</td>
    <td>2.41</td>
    <td><u>73.0</u></td>
    <td><u>86.8</u></td>
    <td><u>91.2</u></td>
    <td><u>95.3</u></td>
    <td><u>69.6</u></td>
    <td><u>84.4</u></td>
    <td><u>94.5</u></td>
    <td><b>98.7</b></td>
    <td><u>89.5</u></td>
    <td><u>94.4</u></td>
  </tr>
  <tr>
    <td>UniPR-3D (ours)</td>
    <td>8.23</td>
    <td><b>74.3</b></td>
    <td><b>87.5</b></td>
    <td><b>91.4</b></td>
    <td><b>96.0</b></td>
    <td><b>76.2</b></td>
    <td><b>87.3</b></td>
    <td><b>94.9</b></td>
    <td>98.1</td>
    <td><b>89.6</b></td>
    <td><b>94.5</b></td>
  </tr>
</table>

##### Sequence matching results

<table>
  <tr>
    <th></th>
    <th colspan="3">MSLS Val</th>
    <th colspan="3">NordLand</th>
    <th colspan="3">Oxford1</th>
    <th colspan="3">Oxford2</th>
  </tr>
  <tr>
    <th>Method</th>
    <th>R@1</th>
    <th>R@5</th>
    <th>R@10</th>
    <th>R@1</th>
    <th>R@5</th>
    <th>R@10</th>
    <th>R@1</th>
    <th>R@5</th>
    <th>R@10</th>
    <th>R@1</th>
    <th>R@5</th>
    <th>R@10</th>
  </tr>
  <tr>
    <td>SeqMatchNet</td>
    <td>65.5</td>
    <td>77.5</td>
    <td>80.3</td>
    <td>56.1</td>
    <td>71.4</td>
    <td>76.9</td>
    <td>36.8</td>
    <td>43.3</td>
    <td>48.3</td>
    <td>27.9</td>
    <td>38.5</td>
    <td>45.3</td>
  </tr>
  <tr>
    <td>SeqVLAD</td>
    <td>89.9</td>
    <td>92.4</td>
    <td>94.1</td>
    <td>65.5</td>
    <td>75.2</td>
    <td>80.0</td>
    <td>58.4</td>
    <td>72.8</td>
    <td>80.8</td>
    <td>19.1</td>
    <td>29.9</td>
    <td>37.3</td>
  </tr>
  <tr>
    <td>CaseVPR</td>
    <td><u>91.2</u></td>
    <td><u>94.1</u></td>
    <td><u>95.0</u></td>
    <td><u>84.1</u></td>
    <td><u>89.9</u></td>
    <td><u>92.2</u></td>
    <td><u>90.5</u></td>
    <td><u>95.2</u></td>
    <td><u>96.5</u></td>
    <td><u>72.8</u></td>
    <td><u>85.8</u></td>
    <td><u>89.9</u></td>
  </tr>
  <tr>
    <td>UniPR-3D (ours)</td>
    <td><b>93.7</b></td>
    <td><b>95.7</b></td>
    <td><b>96.9</b></td>
    <td><b>86.8</b></td>
    <td><b>91.7</b></td>
    <td><b>93.8</b></td>
    <td><b>95.4</b></td>
    <td><b>98.1</b></td>
    <td><b>98.7</b></td>
    <td><b>80.6</b></td>
    <td><b>90.3</b></td>
    <td><b>93.9</b></td>
  </tr>
</table>


## Compute Infrastructure

### Hardware
- NVIDIA RTX 4090

### Software
- Python 3.11.10 + CUDA 12.1  
- Based on [SALAD](https://github.com/serizba/salad) and [VGGT](https://github.com/facebookresearch/vggt)

## Citation

**BibTeX:**
```bibtex
@article{deng2025unipr3d,
  title={UniPR-3D: Towards Universal Visual Place Recognition with 3D Visual Geometry Grounded Transformer},
  author={Deng, Tianchen and Chen, Xun and Li, Ziming and Shen, Hongming and Wang, Danwei and Civera, Javier and Wang, Hesheng},
  journal={arXiv preprint arXiv:2512.21078},
  year={2025}
}
```

**APA:**
Deng, T., Chen, X., Li, Z., Shen, H., Wang, D., Civera, J., & Wang, H. (2025). UniPR-3D: Towards Universal Visual Place Recognition with 3D Visual Geometry Grounded Transformer. *arXiv preprint arXiv:2512.21078*.

## Contact

For questions, pretrained model access, or qualitative comparisons, please contact:  
πŸ“§ **Tianchen Deng** – [[email protected]](mailto:[email protected])

---

> πŸ“Œ **Acknowledgement**: This implementation builds upon [SALAD](https://github.com/serizba/salad) and [VGGT](https://github.com/facebookresearch/vggt). Please cite those works if you use their components.