1. Background
Since the outbreak of COVID-19, people’s daily lives and travel have been significantly affected. Wearing masks has become a basic public health requirement in many countries.
While temperature screening and health code systems are now widely used, there is still a lack of automated solutions for detecting whether people are wearing masks correctly.
Currently, mask compliance in most public areas relies on manual supervision, which is inefficient and inconsistent—especially in crowded environments such as hospitals, factories, transport hubs, and shopping malls.
To address this need, Fuwee Vision has developed an AI-powered Mask Detection System, which automatically identifies individuals not wearing masks and helps maintain public safety and compliance with epidemic prevention requirements.
2. Product Overview
The Fuwee AI Mask Detection System is an intelligent image recognition solution based on real-time video stream analysis.
Using the latest deep learning and big data technologies, the system replaces manual observation—automatically detecting when a person in the camera’s field of view is not wearing a mask.
It works seamlessly with existing surveillance cameras, without requiring any additional hardware.
The system performs real-time monitoring and alerts, offering fast, accurate, and cost-effective management of mask compliance.
3. Workflow
-
Front-end cameras continuously upload video streams to the central server.
-
The AI Mask Detection System analyzes the video in real time.
-
When a person not wearing a mask is detected, the system captures an image, saves the evidence, and sends an alert or alarm notification.
4. Features
-
Automatic Detection: Captures and alerts when an unmasked individual is detected.
-
Long-Distance Recognition: Effective recognition range up to 30 meters.
-
Flexible Deployment: Works in all environments and lighting conditions.
-
Wide Camera Compatibility: Supports both bullet and PTZ cameras.
-
Real-Time Processing: Instant recognition and notification.
-
Recommended View: Best performance when the person faces the camera directly.
5. Technical Specifications
Detection Standards:
-
The system requires that the person’s face width ≥ 64 pixels and the mask area is visible to the human eye.
Detection Range (using 2MP camera):
-
8mm lens: up to 30 meters
-
6mm lens: up to 25 meters
-
4mm lens: up to 20 meters
Performance:
-
Capture rate: ≥ 95%
-
Accuracy: ≥ 90%
Hardware Requirements:
-
Operating System: 64-bit Windows (Win10 or above recommended)
-
CPU: Intel Core i3/i5/i7/i9 (4th generation or newer)
-
RAM: ≥ 4GB
-
No dedicated GPU required
-
Network: Must connect to cameras, no internet connection needed
-
One server supports up to 32 camera channels
-
Camera: Hikvision, Dahua, or other RTSP/RTMP/ONVIF network cameras
-
Resolution: ≥ 2 Megapixels
6. Application Scenarios
The Fuwee AI Mask Detection System can be deployed in a wide range of environments, including:
-
Hospitals and healthcare facilities
-
Factories and industrial workshops
-
Restaurants and hotel kitchens
-
Bus and train stations
-
Shopping malls and public buildings
-
Schools and universities
Its real-time and automated monitoring helps ensure compliance with health and safety protocols, improving both management efficiency and public safety.