File size: 2,489 Bytes
0dc9398 | 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 | ---
pretty_name: CelloCut Benchmark
license: cc-by-nc-4.0
viewer: false
size_categories:
- n<1K
tags:
- 3d
- mesh
- benchmark
- geometry-processing
- shape-completion
- watertight-mesh
---
# CelloCut Benchmark
This repository contains the benchmark data and evaluation scripts used for CelloCut.
It is organized as two paired 3D mesh benchmarks: CelloFill and CelloScan.
## Directory Structure
```text
CelloCut_Benchmark/
|-- CelloFill/
| |-- input/ # 110 input meshes, OBJ format
| `-- gt/ # 110 ground-truth meshes, OBJ format
|-- CelloScan/
| |-- input/ # 140 input meshes, OBJ format
| `-- gt/ # 140 ground-truth meshes, PLY format
`-- evaluation/
|-- geometric_metric/
`-- visual_metric/
```
Files are matched by filename stem. For example, predictions for `CelloScan/input/12.obj`
should be named `12.obj`, `12.ply`, or another supported mesh extension accepted by the
evaluation script.
## Dataset Splits
| Benchmark | Input directory | Ground truth directory | Cases | Input format | Ground truth format |
| --- | --- | --- | ---: | --- | --- |
| CelloFill | `CelloFill/input` | `CelloFill/gt` | 110 | OBJ | OBJ |
| CelloScan | `CelloScan/input` | `CelloScan/gt` | 140 | OBJ | PLY |
## Evaluation
Geometric metrics:
```bash
python evaluation/geometric_metric/evaluate_geometry.py \
--gt-dir CelloScan/gt \
--pred-dir path/to/predictions \
--output-dir path/to/geometric_results
```
Visual metrics:
```bash
python evaluation/visual_metric/evaluate_visual.py \
--gt-dir CelloFill/gt \
--pred-dir path/to/predictions \
--output-dir path/to/visual_results \
--blender blender
```
See `evaluation/README.md` for more details on the reported metrics and optional
dependencies.
## Hugging Face Notes
The benchmark uses mesh files (`.obj` and `.ply`). These are distributed as repository
files and are not expected to be rendered by the Hugging Face Dataset Viewer. Download the
repository files directly, then run the evaluation scripts above.
## License
The benchmark data and evaluation scripts are released under the Creative Commons
Attribution-NonCommercial 4.0 International License. See `LICENSE` for details.
If any included asset is derived from a third-party source with a more restrictive license,
the original source license continues to apply to that asset.
## Citation
If you use this benchmark, please cite the CelloCut paper. The BibTeX entry will be updated
when the arXiv version is available.
|