--- title: "2D Printed Mask and Replay Attack Videos Dataset" description: "A multi-part video dataset for detecting 2D print attacks (printed photos) and replay attacks (faces displayed on screens)." url: "https://unidata.pro/datasets/2d-printed-mask-and-replay-attack-videos-dataset/" date_modified: "2025-11-27T19:54:53+03:00" language: "en-US" --- The dataset includes 26,436 MP4 and MOV videos featuring individuals with 2D printed masks, silicone masks, and replay attacks, enriched with metadata (age, gender, ethnicity) and diverse categories such as video fakes, crowdworker spoofs, zoom replays, cut-out printouts, and in-car recordings, designed for training face recognition, liveness detection, and spoofing detection systems. ## Dataset Structure ### The Numbers Section **Numbered list:** - **Number:** 26 436 — **Text:** Videos ### Tooltips Section **Tooltip items:** - **Name:** Computer Vision - **Name:** Machine learning - **Name:** Security - **Name:** Anti-spoofing - **Name:** Liveness Detection ### Dataset Information **Table with data:** | Characteristic | Data | | --- | --- | | Description | Videos of individuals with printed masks and replay attacks | | Data types | Video | | Tasks | Anti-spoofing, Liveness Detection, Face Recognition | | Number of videos | 26 436 | | Name of categories | Video fakes, tolokers based spoofs (Videos of crowdworkers' faces displayed on computer or phone screens), video zoom, youdo heads (Videos of cut-out color printouts of faces, collected via YouDo), in-car videos and fakes | | Labeling | Metadata (age, gender, ethnicity) | | Gender | Male, Female | **Media Slider:** - **Video on Slayder:** - **Video on Slayder:** **Link to the sample:** [Download sample](https://drive.google.com/drive/folders/1FT-GzFN4D3xYPxIA0qhICpb8vblfddYQ?usp=sharing) ### Technical Specifications **Table with data:** | Characteristic | Data | | --- | --- | | Video extensions | MP4, MOV | **Source and data collection methodology:** Source and collection methodology; Data was collected by a partner of Unidata. ### Dataset Use Cases - Slider **Industry Cards:** - **Industry:** Biometrics and Security — **Title:** Face antispoofing for secure authentication — **Text:** 2D Printed Mask and Replay Attack Videos Dataset provides videos of masked attacks, printed photos, and replay attacks that are essential for building robust anti-spoofing methods. Security providers can train recognition systems to detect face spoofing and distinguish real faces from 2D attacks, ensuring higher accuracy in biometric authentication and liveness detection. - **Industry:** Financial Services — **Title:** Fraud prevention in digital KYC — **Text:** This **replay attack dataset** supports banks and fintech companies in developing detection algorithms that identify presentation attacks during identity verification. By exposing recognition systems to 2D print and replay attack videos, the dataset improves spoofing detection, reducing fraudulent attempts with printed photos and replayed recordings in KYC processes. - **Industry:** AI and Machine Learning Research — **Title:** Benchmarking face recognition models — **Text:** Researchers can use the **2D mask attack dataset** as training data for developing and benchmarking new detection methods. With over 26,000 videos covering diverse attack types, learning models trained on this dataset can improve accuracy in detecting 2D attacks, enhance facial image quality analysis, and strengthen recognition systems against spoofing. - **Industry:** Automotive Industry — **Title:** In-car face recognition testing — **Text:** The dataset includes in-car spoofing videos that simulate real-world conditions for vehicle security systems. Developers can use this **anti-spoofing replay attack dataset** to test detection algorithms against printed attacks, 2D masks, and replayed screens, ensuring that in-car recognition systems can identify presentation attacks and provide reliable driver authentication. ### Fact **FAQs Heading:** FAQs **List of Questions:** - **Question:** How was the data collected? — **Answer:** The dataset was collected through Unidata’s trusted partners using controlled setups with real faces, printed masks, and replay attacks. This ensures consistent image quality and realistic presentation attacks for facial images research. - **Question:** How can this dataset help in spoofing research? — **Answer:** By offering varied attack types such as 2D print, replay attacks, and face spoofs, the dataset provides rich training data for anti-spoofing replay attack datasets. Researchers can benchmark and optimize facial recognition and detection methods against real-world presentation attacks. - **Question:** Why are both printed mask attacks and replay attacks included? — **Answer:** Combining physical printed attacks with digital replay attacks allows developers to build more robust liveness detection systems that generalize across different presentation attack techniques instead of specializing in only one spoof type. - **Question:** Can I request a sample before buying or downloading the dataset? — **Answer:** Yes, Unidata offers dataset samples upon request. This allows researchers to evaluate video quality, labeling accuracy, and the suitability of training data for facial recognition or spoofing detection tasks. - **Question:** How are Unidata datasets licensed? — **Answer:** Unidata datasets follow a dual-licensing model. Free samples are available for trial, testing, and evaluation, while full access to the whole dataset is granted exclusively through purchase. - **Question:** Do Unidata datasets follow GDPR or other data privacy regulations? — **Answer:** Yes. All Unidata datasets comply with GDPR and related international data protection laws. This dataset was created from legally sourced materials, ensuring ethical, lawful, and privacy-compliant data collection for face recognition research. - **Question:** How are Unidata datasets stored? — **Answer:** Unidata securely stores all datasets on AWS cloud infrastructure to ensure scalability, high availability, and data security. The storage system adheres to ISO 27001 and ISO 27701 standards, providing a reliable environment for handling sensitive facial recognition and spoofing detection data. - **Question:** How long does it take to receive the dataset? — **Answer:** After you submit a request, Unidata will contact you to verify your requirements and complete the necessary documentation. Once the agreement and payment are finalized, the 2D mask attack dataset is delivered within 3–10 business days. [Full list of this site's AI-readable pages](https://unidata.pro/llms.txt)