--- title: "Fabric Masks and Disguise Presentation Attack Dataset" description: "High-quality face anti-spoofing dataset containing 9,270 HD and 4K videos of people wearing fabric and cloth face masks. Designed for presentation attack detection and face…" url: "https://unidata.pro/datasets/fabric-masks-and-disguise-presentation-attack-dataset/" date_modified: "2026-06-05T11:45:56+03:00" language: "en-US" --- High-quality face anti-spoofing dataset containing 9,270 HD and 4K videos of people wearing fabric and cloth face masks. Designed for presentation attack detection and face recognition research, this face mask detection dataset includes varied lighting, camera types, and demographic diversity to support robust mask detection and biometric security model training. ## Dataset Structure ### The Numbers Section **Numbered list:** - **Number:** 9 270 — **Text:** Videos - **Number:** 103 — **Text:** people ### Tooltips Section **Tooltip items:** - **Name:** Facial Recognition - **Name:** iBeta - **Name:** Liveness Detection - **Name:** Security - **Name:** Anti-spoofing - **Name:** Computer Vision ### Dataset Information **Table with data:** | Characteristic | Data | | --- | --- | | Description | Videos of people in fabric masks training algorithms to detect biometric hacking attempts. | | Data types | Video | | Tasks | Face recognition, Computer Vision | | Total number of videos | 9 270 | | Total number of people | 103 | | Labeling | Only technical characteristics and metadata (age, gender, ethnicity, glasses, wig, camera, light condition, background) | | Gender | Male(50%), Female(50%) | | Ethnicity | Caucasian, African, Asian | **Media Slider:** - **Video on Slayder:** - **Video on Slayder:** **Link to the sample:** [Download sample](https://drive.google.com/drive/folders/1vndMYAkDekrQwVQXzDNVLHdcqVIKh4qE?usp=sharing) ### Statistics - Charts **Charts with Titles:** - **Shortcode:** [ays_chart id="33"] — **caption above the graph:** Distribution by gender ### Technical Specifications **Table with data:** | Characteristic | Data | | --- | --- | | Video Extensions | Mp4, MOV | | Video Resolutions | Min = 1920х1080, Max = 3840х2160 | | Video Duration | 3-5 second | | Number of background | 5 | | Devices | Samsung, iPhone, Xiaomi, Huawei | **Source and data collection methodology:** Source and collection methodology: Data was collected via crowdsourcing platforms. ### Dataset Use Cases - Slider **Industry Cards:** - **Industry:** Biometric Security and Authentication — **Title:** Improving Face Anti-Spoofing Accuracy — **Text:** This face mask detection dataset supports the development of robust face anti-spoofing algorithms by providing high-quality videos containing people wearing different types of face masks, including cloth masks, cotton masks, and surgical masks. The dataset consists of varied backgrounds, lighting conditions, and age groups, enabling trained models to reliably detect presentation attacks and improve face recognition performance across real-world scenarios. - **Industry:** Healthcare and Public Safety — **Title:** Enhancing Mask Detection for Compliance Monitoring — **Text:** This fabric mask dataset enables AI systems to detect and monitor mask usage in public and healthcare settings. The dataset’s videos include medical masks, cloth masks, and other mask types, recorded under multiple lighting conditions and backgrounds. Such datasets containing real-world mask-wearing behavior allow learning algorithms to achieve higher recognition accuracy, supporting safe operations in hospitals, public transport, and crowded environments. - **Industry:** AI Research and Machine Learning — **Title:** Benchmarking Recognition Models Under Masked Conditions — **Text:** Researchers can leverage this face anti-spoofing dataset to study the impact of mask types and occlusions on face recognition systems. The dataset contains annotated metadata for gender, ethnicity, and lighting, allowing trained models to evaluate results obtained across different age groups. This enhances learning algorithms and provides a solid foundation for academic research on presentation attack resilience. - **Industry:** Consumer Electronics and Smart Devices — **Title:** Optimizing Biometric Systems for Masked Users — **Text:** This face anti-spoofing dataset is critical for testing smartphone and camera-based recognition systems. It includes recordings from devices like Samsung, Huawei, Xiaomi, and iPhone, covering multiple mask types and backgrounds. Developers can use this dataset to train models that maintain face recognition accuracy under mask usage, ensuring reliable biometric authentication in everyday applications. ### Fact **FAQs Heading:** FAQs **List of Questions:** - **Question:** What is included in this dataset? — **Answer:** The dataset comprises 9 270 high-resolution video clips of 103 individuals wearing various fabric masks and disguises. Videos include metadata such as age, gender, ethnicity, glasses, wigs, camera type, lighting, and background, with video durations of 3–5 seconds. - **Question:** What types of annotations are provided? — **Answer:** Each video is labeled with technical metadata, including subject age, gender, ethnicity, mask type, accessories, lighting conditions, camera device, and background. These annotations enable supervised learning for face anti-spoofing and mask detection systems. - **Question:** Can I request a sample of the dataset before purchasing or downloading it? — **Answer:** Yes. You can request a sample of the dataset to verify video quality, mask diversity, and metadata completeness. Sampling ensures that the dataset meets your requirements for training anti-spoofing models or conducting face mask detection research. - **Question:** What are the sources of data for Unidata datasets? — **Answer:** Unidata datasets are collected from verified partners and legally compliant sources. The Fabric Masks Dataset was collected via crowdsourcing platforms, capturing videos of people using different fabric masks, wigs, and glasses. - **Question:** How was the dataset collected? — **Answer:** Data was collected via crowdsourcing platforms, recording individuals in controlled indoor and outdoor environments using devices like Samsung, iPhone, Xiaomi, and Huawei smartphones. The dataset covers five different backgrounds and various lighting conditions to simulate realistic presentation attacks. - **Question:** Do Unidata datasets follow GDPR or other data privacy regulations? — **Answer:** Yes. All Unidata datasets, including this fabric masks and disguise dataset, comply with GDPR and applicable data protection laws. Data is collected from legally permissible sources, ensuring ethical and lawful usage. - **Question:** How are Unidata datasets stored? — **Answer:** Unidata stores all datasets on AWS cloud infrastructure, ensuring high availability, scalability, and secure data management. Storage practices follow ISO 27001 and ISO 27701 standards, providing a privacy-focused environment for sensitive biometric data. - **Question:** How long does it take to receive the dataset? — **Answer:** Once your request is submitted, Unidata reviews the details and completes the necessary documentation. After signing and payment, the dataset is delivered within 3–10 business days. [Full list of this site's AI-readable pages](https://unidata.pro/llms.txt)