codershiyar/imdb-clean-2010plus-singleface

IMDB-Clean 2010+ SingleFace Subset

Dataset Description

This dataset is a filtered subset of IMDB-Clean for age estimation and resolution impact studies.
It contains 1,000 face images selected under strict criteria:

• Source: IMDB-WIKI / IMDB-Clean
• Year filter: only photos taken after 2010
• Face detection: exactly one face per image (using InsightFace)
• Image quality: resolution ≥ 1024 px

A metadata CSV (subset_metadata.csv) is included with age, bounding box, and resolution info.

Preprocessing

• Original images filtered and copied into filtered_imdb_clean_images
• Resized versions generated via letterbox padding into multiple resolutions:
  • 64, 112, 224, 512, 720, 1080, 2160 px

These are stored in the resized_images folder.

Dataset Statistics

• Original filtered images: 1,000
• Resized versions: 8,000 (8 per original)
• Total images in repository: ~9,000

Intended Uses

• Benchmarking age estimation frameworks (DeepFace, InsightFace, etc.)
• Studying the effect of input resolution on inference time and accuracy
• Teaching / reproducible experiments in AI coursework

⚠️ Note: This dataset is intended strictly for research and educational use.
It should not be used for commercial applications or biometric identification.

Limitations

• Dataset size is limited (1,000 images).
• Labels come from IMDB metadata → possible noise.
• Some demographic bias inherited from IMDB-WIKI.

Creator and Maintainer

This subset was curated and prepared by Shiyar Jamo
(Master’s in Artificial Intelligence, Amsterdam University of Applied Sciences, 2025).

If you use this dataset, please also cite this repository in addition to the original sources.

Citation

If you use this dataset, please cite both the original IMDB-WIKI/IMDB-Clean dataset and this repository:

@inproceedings{rothe2015dex,
  title={DEX: Deep expectation of apparent age from a single image},
  author={Rothe, Rasmus and Timofte, Radu and Van Gool, Luc},
  booktitle={Proceedings of the IEEE international conference on computer vision workshops},
  year={2015}
}

@inproceedings{lin2021fp,
  title={FP-Age: Leveraging Face Parsing Attention for Facial Age Estimation in the Wild},
  author={Lin, Yijun and Shen, Jiayi and Wang, Yu and Pantic, Maja},
  booktitle={Proceedings of the IEEE Winter Conference on Applications of Computer Vision},
  pages={2060--2069},
  year={2021}
}

@misc{jamo2025imdbclean,
  title={IMDB-Clean 2010+ SingleFace Subset},
  author={Jamo, Shiyar},
  year={2025},
  howpublished={Hugging Face dataset},
  url={https://huggingface.co/datasets/codershiyar/imdb-clean-2010plus-singleface}
}