TaxonomyProject/SimulationMetadata

1---2license: cc-by-4.03task_categories:4- visual-question-answering5- image-classification6tags:7- 3d8- spatial-reasoning9- object-properties10- taxonomy11- simulation12pretty_name: PercepTax Simulation Metadata13---14 15# PercepTax — Simulation Metadata16 17Supporting metadata for the **PercepTax** benchmark.18 19<p align="center">20  <a href="https://perceptual-taxonomy.github.io"><img alt="Project Page" src="https://img.shields.io/badge/Project_Page-2563EB?style=for-the-badge&logo=googlechrome&logoColor=white"></a>21  <a href="https://arxiv.org/abs/2511.19526"><img alt="Paper" src="https://img.shields.io/badge/arXiv-2511.19526-B31B1B?style=for-the-badge&logo=arxiv&logoColor=white"></a>22  <a href="https://github.com/XingruiWang/PercepTaxBench"><img alt="Code" src="https://img.shields.io/badge/Code-181717?style=for-the-badge&logo=github&logoColor=white"></a>23  <a href="https://huggingface.co/datasets/RyanWW/PercepTaxBench"><img alt="VQA Dataset" src="https://img.shields.io/badge/VQA_Dataset-FFD21E?style=for-the-badge&logo=huggingface&logoColor=black"></a>24</p>25 26This repository holds the **object taxonomy**, **object lists / placement rules**, and27**per-scene 3D annotations** used to render and annotate the simulated split of PercepTax.28It is a collection of JSON/CSV/PNG files (not a `load_dataset`-style tabular dataset).29 30## Contents31 32```33taxonomy/                     # the physical-property taxonomy34├── taxonomy.json             # full taxonomy tree35├── all_objects.json          # object → property assignments36├── material.csv  shape.csv  color.csv  texture.csv37├── affordance.csv  function.csv  environment.csv  physical_properties.csv38 39objects/                      # object catalog & placement rules40├── background_objects.json41├── objects_list/             # object lists & class↔instance mappings42└── object_placement/         # placable-on-table / -chair / -floor lists43 44scenes/                       # per-scene 3D annotations45├── annotations/<SceneName>/<view_id>/46│   ├── scene_annotations_split.json   # camera, background, foreground objects,47│   │                                  #   3D boxes, image_path48│   ├── bbox_visualization_all.png     # 2D box overlay49│   └── bbox_3d_visualization.png      # 3D box overlay50├── filter/                            # filtered scene subsets51└── image_quality_ratings*.json        # automated scene quality scores52```53 54### `scene_annotations_split.json` fields55`scene_name`, `view_id`, `image_path`, `camera` (intrinsics/extrinsics),56`background`, `foreground` (objects with labels and 3D bounding boxes).57 58## Usage59 60Browse or download with the Hub API:61 62```python63from huggingface_hub import snapshot_download64path = snapshot_download("TaxonomyProject/SimulationMetadata", repo_type="dataset")65```66 67The rendered scene **images** for the simulated VQA split are bundled in68[`RyanWW/PercepTaxBench`](https://huggingface.co/datasets/RyanWW/PercepTaxBench) (config `sim`).69 70## License71 72[CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/).73