---
title: "Silicone Mask Attack Dataset"
description: "6,500+ videos 5 devices"
url: "https://unidata.pro/datasets/silicone-mask-attacks/"
date_modified: "2025-12-09T09:03:20+03:00"
language: "en-US"
---
The dataset contains videos of human faces with silicone masks designed for presentation attacks, providing high-quality data for face recognition, anti-spoofing, and attack detection research, supporting the development of robust recognition systems, detection algorithms, and deep learning models against mask-based biometric attacks in compliance with iBeta Level 2 certification standards

## Dataset Structure

### The Numbers Section

**Numbered list:**

- **Number:** 6,500+ — **Text:** Videos
- **Number:** 5 — **Text:** Devices

### Tooltips Section

**Tooltip items:**

- **Name:** iBeta
- **Name:** Liveness Detection
- **Name:** Computer Vision
- **Name:** Security
- **Name:** Facial Recognition
- **Name:** Anti-spoofing

### Dataset Information

**Table with data:**

| Characteristic | Data |
| --- | --- |
| Description | Videos of people in silicone masks training algorithms to detect biometric hacking attempts. |
| Data types | Video |
| Tasks | Face recognition, Computer Vision |
| Total number of videos | 6,500 |
| Total number of people | 50 |
| Labeling | Only technical characteristics and metadata (age, gender, ethnicity) |
| Gender | Male, Female |
| Ethnicity | Caucasian (90%), African (10%) |
| Number of attributes | 31 |

**Media Slider:**

- **Video on Slayder:** [https://unidata.pro/wp-content/uploads/2024/10/silicone\_1.mp4](https://unidata.pro/wp-content/uploads/2024/10/silicone_1.mp4)
- **Video on Slayder:** [https://unidata.pro/wp-content/uploads/2024/10/silicone\_2.mp4](https://unidata.pro/wp-content/uploads/2024/10/silicone_2.mp4)

**Link to the sample:** [Download sample](https://drive.google.com/drive/folders/1S4bnkyKl9X5B7RreEJ2Ci7VbZBZF19WQ)

### Technical Specifications

**Table with data:**

| Characteristic | Data |
| --- | --- |
| Video Extensions | mp4 |
| Video Resolutions | Min = 1920х1080, Max = 3840х2160 |
| Video Duration | 1-2 second |
| Number of background | 9 |
| Video Overlap | No more than 5% |
| Devices | Mi10s, Google Pixel 4, Samsung Galaxy A03s, iPhone 11, iPhone SE 2 and etc. |

**Source and data collection methodology:** Source and collection methodology:  Data was collected via crowdsourcing platforms.

### Statistics - Charts

**Charts with Titles:**

- **Shortcode:** [ays_chart id='34'] — **caption above the graph:** Number of devices of each type
- **Shortcode:** [ays_chart id='35'] — **caption above the graph:** Ethnicity in the dataset
- **Shortcode:** [ays_chart id='36'] — **caption above the graph:** Gender distribution

### Dataset Use Cases - Slider

**Industry Cards:**

- **Industry:** Biometrics and Security — **Title:** Developing silicone mask anti-spoofing systems — **Text:** Silicone Mask Attack Dataset provides realistic recordings of silicone masks, latex masks, and other presentation attacks. Recognition systems trained on this dataset can accurately detect spoofing attacks, improving biometric authentication and ensuring stronger protection in face recognition applications.
- **Industry:** Financial Services — **Title:** Preventing identity fraud in KYC — **Text:** Banks and fintech companies use mask attack datasets to train recognition algorithms that detect 3D masks and other face presentation attacks. Learning models built on this data enhance liveness detection during onboarding, reducing risks of fraudulent verification through silicone masks or replay attacks.
- **Industry:** AI and Machine Learning Research — **Title:** Benchmarking detection algorithms — **Text:** This dataset serves as training data for deep learning and image classification research. With diverse facial features and attack types, it allows researchers to compare recognition algorithms, test detection methods, and improve deep models for spoofing attack detection.
- **Industry:** Forensics and Law Enforcement — **Title:** Enhancing facial recognition in investigations — **Text:** Law enforcement agencies can use Silicone Mask Attack Dataset to train recognition systems against synthetic dataset disguises. The database contains 3D masks, paper masks, and silicone masks, providing realistic cases for developing detection algorithms capable of identifying real faces hidden by advanced disguises during forensic analysis.

### Fact

**FAQs Heading:** FAQs

**List of Questions:**

- **Question:** How was Silicone Mask Attack Dataset collected? — **Answer:** The dataset was collected through crowdsourcing platforms, using devices such as Mi10s, Google Pixel 4, and iPhone 11. This method ensures a large collection of high-resolution facial images across various ages and backgrounds.
- **Question:** What are the technical characteristics of the dataset? — **Answer:** The dataset consists of MP4 videos with resolutions ranging from 1920×1080 to 3840×2160. Each video lasts 1–2 seconds, with up to 9 unique backgrounds, providing diverse training data for image classification and deep learning research.
- **Question:** Does the dataset contain both real and masked faces? — **Answer:** Yes, the dataset includes real faces and silicone masks to simulate mask attacks. This combination helps researchers build robust detection methods capable of distinguishing the facial features of synthetic datasets from authentic human faces.
- **Question:** How does this dataset support spoofing detection research? — **Answer:** By including silicone masks and varied spoofing attacks, the dataset enables testing of deep learning algorithms for anti-spoofing. It is particularly useful for evaluating recognition systems in high-security scenarios where presentation attacks are a concern.
- **Question:** Do Unidata datasets follow GDPR or other data privacy regulations? — **Answer:** Yes. All Unidata datasets comply with GDPR and relevant international data privacy regulations. This dataset was curated from legally permissible sources to support ethical research in face presentation and biometric attack detection without compromising personal information or privacy standards.
- **Question:** How are Unidata datasets stored? — **Answer:** Unidata securely stores all datasets on AWS cloud infrastructure. Storage and management practices meet ISO 27001 and ISO 27701 standards, ensuring this Silicone Mask Dataset is protected within a privacy-focused environment optimized for reliability, scalability, and data integrity.
- **Question:** How long does it take to receive the dataset? — **Answer:** After submitting a request, Unidata contacts you to confirm details and complete required documentation. Once the agreement and payment are finalized, a dataset is usually delivered within 3–10 days.
- **Question:** Is this a real-world dataset or synthetic data? — **Answer:** This is a real-world dataset collected through crowdsourcing platforms featuring genuine presentation attacks using silicone masks. Unlike synthetic datasets or simulated print attacks, it captures authentic mask attacks on facial recognition systems, offering realistic training data for deep models and object detection algorithms.
- **Question:** Can the dataset be used for deep learning and computer vision applications? — **Answer:** Yes. The combination of high-resolution videos, realistic silicone mask attacks, diverse recording devices, multiple backgrounds, and detailed metadata makes the dataset suitable for training deep learning models for face recognition, presentation attack detection, liveness detection, and computer vision applications.

[Full list of this site's AI-readable pages](https://unidata.pro/llms.txt)