---
license: cc-by-4.0
task_categories:
- image-feature-extraction
- image-classification
- video-classification
language:
- en
tags:
- liveness detection
- anti-spoofing
- biometrics
- facial recognition
- machine learning
- deep learning
- AI
- paper mask attack
- iBeta certification
- PAD attack
- security
- ibeta
- face recognition
- pad
- authentication
- fraud
---
# Liveness Detection Dataset: 3D Paper Mask Attacks

Paper mask attack dataset for training PAD and liveness detection models against low-cost 3D spoofing — 2,000+ videos on iOS and Android, ISO 30107-3 Level 1/2 attack category

## What This Dataset Covers

2,000+ video recordings of 3D paper mask presentation attacks — masks with volumetric elements that simulate facial depth. Captured on iOS and Android devices, ~7 sec per video, with zoom-in/zoom-out phases for active liveness scenarios

- 40+ participants
- Balanced gender and ethnicity representation (Caucasian, Black, Asian)
- Multiple device types: low-end to flagship

## Attack Classification (ISO 30107-3)

This attack type falls under **3D artifact attacks** in the ISO 30107-3 taxonomy — the same category tested in iBeta Level 1 and Level 2 certification. Paper masks with volume elements (protruding nose, forehead, shoulders) are one of the most accessible attack vectors, making them a baseline requirement for any PAD system
![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F20109613%2Fc1a66077f517a6165a102023b369e0c1%2FMerged.jpg?generation=1730402472776480&alt=media)

## How This Dataset Fits into Anti-Spoofing Pipelines

- **Standalone:** Train a binary classifier (live vs. paper mask attack)
- **Combined:** Merge with other attack types for a multi-class PAD model — pairs well with replay, print, and silicone mask data
- **iBeta prep:** Use alongside live session data to simulate Level 1 certification test conditions


## Related Datasets

For a more advanced version of this attack type, see:  
[**Wrapped 3D Paper Mask Spoofing Dataset**](https://huggingface.co/datasets/AxonData/Wrapped_3D_Attacks) — full-face masks 
on mannequins with wigs and accessories

For the complete anti-spoofing collection:  
[**Face Anti-Spoofing Dataset**](https://huggingface.co/datasets/AxonData/face-anti-spoofing-dataset) — 100,000+ videos, 11 attack types.

## Full version of the dataset is available for commercial usage. Leave a request on our website [Axonlabs](https://axonlab.ai/?utm_source=hugging-face&utm_medium=cpc&utm_campaign=profile&utm_content=profile_link) to purchase the dataset 💰

## Keywords
- iBeta Certifications
- PAD Attacks
- Presentation Attack Detection
- Antispoofing
- Liveness Detection
- Spoof Detection
- Facial Recognition
- Biometric Authentication
- Security Systems
- AI Dataset
- 3D Mask Attack Dataset
- Active Liveness
- Anti-Spoofing Technology
- Facial Biometrics
- Machine Learning Dataset
- Deep Learning

## Contact and Feedback
We welcome your feedback! Feel free to reach out to us and share your experience with this dataset. If you're interested, you can also **receive additional samples for free**! 😊

Visit us at [**Axonlabs**](https://axonlab.ai/?utm_source=hugging-face&utm_medium=cpc&utm_campaign=profile&utm_content=profile_link) to request a full version of the dataset for commercial usage.