FairFace_Balanced_3K

Overview

FairFace_Balanced_3K is a balanced subset of the original HuggingFaceM4/FairFace dataset created to support bias-sensitive experiments in facial attribute recognition. This subset includes 3,031 samples, with 433 images per race class, across 7 race categories:

• White
• Black
• East Asian
• Southeast Asian
• Indian
• Middle Eastern
• Latino_Hispanic

Each entry contains:

• RGB facial image
• Age group label (9 categories)
• Gender label (Male, Female)
• Race label (7 classes)

The subset is balanced to mitigate data bias and allow fair evaluation across racial groups.

Some random samples of the dataset are as:

Data Format

The dataset is stored in the Hugging Face Hub using the datasets library and Parquet format. Each row includes:

• image: PIL image in byte format
• age: Integer (mapped to age group)
• gender: Integer (0: Male, 1: Female)
• race: Integer (0–6, mapped to race category)

Visuals

• Race Distribution

  

• Age vs Group Distriution

  

• Gender Distribution

  

Citation

If you use this dataset or the original FairFace dataset, please cite the following work:

@inproceedings{karkkainenfairface,
  title={FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age for Bias Measurement and Mitigation},
  author={Karkkainen, Kimmo and Joo, Jungseock},
  booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
  year={2021},
  pages={1548--1558}
}

Original Sources

Original Repo: joojs/fairface

Original Dataset: HuggingFaceM4/FairFace

Original Paper: FairFace