Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# SparseCraft
|
| 2 |
+
|
| 3 |
+
## [ECCV'24] SparseCraft: Few-Shot Neural Reconstruction through Stereopsis Guided Geometric Linearization
|
| 4 |
+
## [Project](https://sparsecraft.github.io/) | [Paper](https://arxiv.org/abs/2407.14257) | [Code](https://github.com/maeyounes/SparseCraft)
|
| 5 |
+
### DTU Dataset
|
| 6 |
+
We provide preprocessed DTU data and results for the tasks of novel view synthesis and surface reconstruction.
|
| 7 |
+
|
| 8 |
+
It contains the following directories:
|
| 9 |
+
|
| 10 |
+
```
|
| 11 |
+
sparsecraft_data
|
| 12 |
+
├── nvs # Novel View Synthesis task data and results
|
| 13 |
+
│ └── mvs_data
|
| 14 |
+
│ ├── scan103
|
| 15 |
+
│ ├── ...
|
| 16 |
+
│ └── results # Results for training using 3, 6, and 9 views
|
| 17 |
+
│ ├── 3v
|
| 18 |
+
│ │ ├── scan103
|
| 19 |
+
│ │ ├── ...
|
| 20 |
+
│ ├── 6v
|
| 21 |
+
│ │ ├── scan103
|
| 22 |
+
│ │ ├── ...
|
| 23 |
+
│ └── 9v
|
| 24 |
+
│ ├── scan103
|
| 25 |
+
│ ├── ...
|
| 26 |
+
└── reconstruction # Surface Reconstruction task data and results
|
| 27 |
+
└── mvs_data # Surface reconstruction data uses a different set of scans and views than the novel view synthesis task
|
| 28 |
+
├── set0
|
| 29 |
+
│ ├── scan105
|
| 30 |
+
│ ├── ...
|
| 31 |
+
└── set1
|
| 32 |
+
├── scan105
|
| 33 |
+
├── ...
|
| 34 |
+
└── results
|
| 35 |
+
├── set0
|
| 36 |
+
│ ├── scan105
|
| 37 |
+
│ ├── ...
|
| 38 |
+
└── set1
|
| 39 |
+
├── scan105
|
| 40 |
+
```
|
| 41 |
+
|
| 42 |
+
**Note**
|
| 43 |
+
|
| 44 |
+
The DTU dataset was preprocessed as follows:
|
| 45 |
+
- The original data is from the [NeuS Project](https://github.com/Totoro97/NeuS). We use the same camera poses and intrinsics as the original data.
|
| 46 |
+
- To obtain MVS data, we used the [Colmap](https://colmap.github.io/) initialized with the original camera poses and intrinsics.
|
| 47 |
+
- We provide a script that achieves this in `scripts` that you can run using the following command. Note that you will need to have Colmap installed on your machine:
|