MaskCognizer

AI-based Face Mask Detection for Public Health and Safety

The AI-based MaskCognizer, developed by Fraunhofer IDMT, is a technology that detects and verifies in real time whether people are wearing their face masks correctly. It provides immediate and appropriate feedback for the protection of the general public and enables easier compliance with public health and safety regulations.

Verification in Motion

Fraunhofer IDMT's robust video analysis technology, embedded in any MaskCognizer-enabled device, classifies up to ten faces simultaneously. It works even when people are in motion. It is not even necessary that a person is positioned in front of the device. Verification with corresponding feedback takes place as the person passes by.

Public Health Safeguards and Cost Savings 

With MaskCognizer-enabled solutions, mask wearing compliance can be checked automatically. This reduces the risk of infection for employees and reduces the security personnel costs associated with monitoring face mask obligations.

Application Areas

Solutions based on the MaskCognizer technology can be deployed wherever mask wearing compliance must be monitored, e.g. in retail, public transport, hospitals, doctors' surgeries, in public administration or in the catering industry, and these are just some of the use cases.

The video shows a prototype of the MaskCognizer. In addition to the Ilmenau video analysis technology it includes intelligent audio system technologyfrom the Oldenburg Branch Hearing, Speech and Audio Technology HSA. An integrated voice output can provide an acoustic cue in addition to the visual feedback.

Benefits

  • Automated detection of masked and unmasked faces
  • Immediate classification of whether the mask is worn correctly (green), or incorrectly/no mask 
  • Multi-person detection: up to ten faces can be classified simultaneously 
  • Ease of use: no need to position individuals directly in front of the device
  • Precise real-time monitoring and feedback to assure compliance with public health and safety requirements
  • Privacy compliant: no storage of personal data
  • Individual feedback options can be implemented, e.g. visual feedback via display, customer-configurable sound or voice output
  • External control options, e.g. to switch light displays or open doors 

Technical Data

  • Supported image sizes: 1280x720 or 1920x1080 (recommended)
  • Recommended camera position: mounting height approx. 1.6 m, preferably frontal view of the face (+/-15°), no back light
  • Real-time detection: detection rate up to 10 frames/sec (Jetson Nano)
  • Mask models are interchangeable and updateable 
  • Example application: camera with horizontal angle of view of 66°, max. distance approx. 5 m, width of monitored area approx. 6.5 m

Business Unit Media Management and Delivery

More information on products and services of the business unit Acoustics