COVID Analytics™ - AI based Intelligent Analytics Solutions

Face Mask Detector Using Artificial Intelligence and Deep Learning

The face mask detector we build, potentially aims at ensuring public safety. Face mask detection offers several advantages to the monitoring personnel. These include face recognition and intelligent alerts. Face mask detection at public places shall ensure public health and safety. People without a mask can be easily identified and notified to wear mask before entering places like malls or transit areas like airports, train stations, etc. Face mask detectors are useful in enterprises too. They ensure whether the workforce maintains safety standards at work.

Our face mask detection system uses AI as a tool based on deep learning algorithms. It detects face masks in public places by performing detection in real time with the help of surveillance camera network. The mask detection system detects face mask in images as well as in live video streams.

Using a detector and classifier for face mask detection

As shown in the block diagram, we get to the results o mask detection in two stages.

Detecting faces using detector- We use trained detector for detecting faces in the stream. The system uses deep convolution neural network-based learning models. We use deep learning classifiers or DL algorithms to extract features and then we use trained classifiers to test the real time data for face mask detection.
Detecting mask on the face ROI- We run trained classifiers on the faces detected by the detector to check whether the faces detected in the stream have their respective masks. The trained model achieves high accuracy with negligible false alarms.

Request Brochure

COVID Analytics™ - AI based Intelligent Analytics Solutions

Simple method to detect Mask or No Mask

This is an easy method to achieve the result. The system is fed with faces, torso, hands and other parts of the body. The landmarks help the system to automatically locate the face. The system then detects the region of interest (RoI) in the faces where mask is worn. The face mask is detected in three simple steps:

Extracting face RoI
Pre-processing RoI
Performing mask detection to check presence of mask or no mask

The system is equipped to detect masks in images and in video streams and is capable of working with minimum computational capability while processing real time image data. It is an easy method to manage wearing of face mask in public places, specially the crowded ones.

The face mask detector can be integrated with the embedded systems and easily used at transit stations, airports, offices and other public places thus ensuring public safety.Block diagram depicting steps to build face mask detector with deep learning and computer vision algorithms is shown.