Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
## Datasets
|
| 2 |
+
1. We conduct experiments on three new 3D domain generalization ([3DDG](#new-3ddg-benchmarks)) benchmarks proposed by us, as introduced in the next section.
|
| 3 |
+
- base-to-new class generalization (base2new)
|
| 4 |
+
- cross-dataset generalization (xset)
|
| 5 |
+
- few-shot generalization (fewshot)
|
| 6 |
+
|
| 7 |
+
2. The structure of these benchmarks should be organized as follows.
|
| 8 |
+
```sh
|
| 9 |
+
/path/to/Point-PRC
|
| 10 |
+
|----data # placed in the same level of `trainers`, `weights`, etc.
|
| 11 |
+
|----base2new
|
| 12 |
+
|----modelnet40
|
| 13 |
+
|----scanobjectnn
|
| 14 |
+
|----shapenetcorev2
|
| 15 |
+
|----xset
|
| 16 |
+
|----corruption
|
| 17 |
+
|----dg
|
| 18 |
+
|----sim2real
|
| 19 |
+
|----pointda
|
| 20 |
+
|----fewshot
|
| 21 |
+
|----modelnet40
|
| 22 |
+
|----scanobjectnn
|
| 23 |
+
|----shapenetcorev2
|
| 24 |
+
```
|
| 25 |
+
|
| 26 |
+
3. You can find the usage instructions and download links of these new 3DDG benchmarks in the following section.
|
| 27 |
+
|
| 28 |
+
## New 3DDG Benchmarks
|
| 29 |
+
### _Base-to-new Class Generalization_
|
| 30 |
+
1. The datasets used in this benchmark can be downloaded according to the following links.
|
| 31 |
+
- [ModelNet40](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/base2new/modelnet40)
|
| 32 |
+
- [S-OBJ_ONLY](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/base2new/scanobjectnn/obj_only)
|
| 33 |
+
- [S-OBJ_BG](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/base2new/scanobjectnn/obj_bg)
|
| 34 |
+
- [S-PB_T50_RS](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/base2new/scanobjectnn/hardest)
|
| 35 |
+
- [ShapeNetCoreV2](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/base2new/shapenetcorev2)
|
| 36 |
+
|
| 37 |
+
2. The following table shows the statistics of this benchmark.
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
### _Cross-dataset Generalization_
|
| 42 |
+
1. The datasets used in this benchmark can be downloaded according to the following links.
|
| 43 |
+
- [OOD Generalization](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/xset/dg)
|
| 44 |
+
- OmniObject3d (Omin3D)
|
| 45 |
+
- [Data Corruption](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/xset/corruption)
|
| 46 |
+
- ModelNet-C (7 types of corruptions)
|
| 47 |
+
- add global outliers, add local outliers, dropout global structure, dropout local region, rotation, scaling, jittering
|
| 48 |
+
- [Sim-to-Real](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/xset/sim2real)
|
| 49 |
+
- [PointDA](https://huggingface.co/datasets/auniquesun/Point-PRC/tree/main/new-3ddg-benchmarks/xset/pointda)
|
| 50 |
+
|
| 51 |
+
2. The following table shows the statistics of this benchmark.
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### _Few-shot Generalization_
|
| 56 |
+
1. Although this benchmark contains same datasets as the _Base-to-new Class_, it investigates the model generalization under extremely low-data regime (1, 2, 4, 8, and 16 shots), which is quite different from the evaluation setting in _Base-to-new Class Generalization_.
|
| 57 |
+
|
| 58 |
+
2. The following table shows the statistics of this benchmark.
|
| 59 |
+
|
| 60 |
+
|