chaofanma/Fantastic-Beasts

Fantastic Beasts Datasets: Benchmark in AttrSeg: Open-Vocabulary Semantic Segmentation via Attribute Decomposition-Aggregation

🤗 View on Hugging Face | 💻 GitHub Repository

This repository contains the collected dataset used in the NeurIPS 2023 paper: AttrSeg: Open-Vocabulary Semantic Segmentation via Attribute Decomposition-Aggregation. See the paper here.

Brief Introduction

Existing datasets often lack the inclusion of rare or obscure vocabulary. To address this limitation, we manually curated a dataset titled "Fantastic Beasts", which consists of 20 categories of magical creatures from the film series Fantastic Beasts and Where to Find Them. This dataset is designed for comprehensive evaluation and simulating real-world scenarios, specifically for two common situations where attribute descriptions are essential:

Neologisms: Vanilla category names represent new vocabularies that are often unseen by large language models (LLMs) and vision-language pre-trainings (VLPs).

Unnameability: When users encounter unfamiliar objects, they may struggle to name them, particularly in the case of rare or obscure categories.

For more details, please refer to the paper.

How to Use This Dataset

Using Hugging Face Datasets

from datasets import load_dataset, Image

# Option 1: Load from Hugging Face Hub
dataset = load_dataset("chaofanma/Fantastic-Beasts", split='test')

# Option 2: Load from local Parquet file
# dataset = load_dataset('parquet', data_files={'test': 'data/test-00000-of-00001.parquet'}, split='test')

# Use the dataset
sample = dataset[0]
image = sample['image']          # PIL Image (embedded in Parquet)
mask = sample['mask']            # PIL Image (embedded in Parquet)
category = sample['category']    # str: "Augurey"
attributes = sample['attributes'] # list of str

Note:

• This dataset uses Parquet format with embedded images (recommended by Hugging Face)
• Images are stored as binary data within the Parquet file (79.8 MB)
• The file test-00000-of-00001.parquet is automatically recognized as the test split
• No manual configuration needed for the Dataset Viewer
• Contains 251 samples designed for evaluation and benchmarking

Alternative: PyTorch Dataset & Raw Images

This Hugging Face repository contains the dataset in Parquet/Arrow format for easy loading.

For alternative formats and implementations, please visit the GitHub Repository which includes:

• Custom PyTorch Dataset class (examples/fantastic_beasts_dataset.py)
• Source images and masks in original quality (organized by category)
• JSONL metadata files
• Additional example scripts

The examples/ folder in this repo provides a reference implementation for those who need it.

Dataset Structure

Category Names and Attributes

There are 20 categories in Fantastic Beasts dataset, listed as below in alphabetical order:

Augurey, Billywig, Chupacabra, Diricawl, Doxy, Erumpent, Fwooper, Graphorn, Grindylow, Kappa, Leucrotta, Matagot, Mooncalf, Murtlap, Nundu, Occamy, Runespoor, Swoopingevil, Thunderbird, Zouwu

Dataset Files

On Hugging Face Hub:

• data/test-00000-of-00001.parquet: Parquet file with embedded images (79.8 MB, contains all 251 samples)
  • Images and masks stored as binary data (Apache Arrow format)
  • Automatically recognized as test split
  • Optimized for streaming and fast loading

For source images and additional formats (JSONL, PyTorch Dataset), see GitHub

Data Fields

• image: PIL Image of the magical creature (embedded binary data in Parquet)
• mask: PIL Image of binary segmentation mask (embedded binary data in Parquet; 0 for background, 255 for object)
• category: Category name (one of 20 magical creature types)
• attributes: List of textual attribute descriptions for the category

Citation

If this dataset is useful for your research, please consider citing:

@article{ma2023attrseg,
  title   = {AttrSeg: Open-Vocabulary Semantic Segmentation via Attribute Decomposition-Aggregation},
  author  = {Chaofan Ma and Yuhuan Yang and Chen Ju and Fei Zhang and Ya Zhang and Yanfeng Wang},
  journal = {Thirty-seventh Conference on Neural Information Processing Systems (NeurIPS)},
  year    = {2023}
}

Acknowledgements

We would like to thank the following people for their direct or indirect contributions to the creation of this dataset:

• J.K. Rowling, as the creator of the Wizarding World and the original author of the Harry Potter series, whose work is foundational.
• David Yates, the director of the film, for contributing to its vision and execution.
• David Heyman, the producer of the film, for his pivotal role in bringing the story to the screen.
• The VFX artists and technicians at Framestore and their team leaders, Tim Burke, Christian Manz, and Pablo Grillo, for their incredible work in creating the magical creatures.
• All the Harry Potter fans who support me in creating this dataset.