Skip to content
Portfolio

Face Mask Detection

Shopper in a grocery store with mask-detection boxes labeled MASK DETECTED and NO MASK

Overview

We created a Face Mask detection system combining the best of Facial recognition and computer vision.

The system tracks people, analyzes faces and generates alerts if necessary. A state of the art Facial recognition system with real time performance, we implemented it in C++ to guarantee stable and seamless deployability. It can work on multiple cameras simultaneously making it equally ideal for small businesses, public sectors (healthcare, transport, schools etc.) and large enterprises with use cases like smart cities.

Mask Detection System

Problem

We were approached to create a system that would detect if people were wearing face masks properly. The system had to work through a CCTV camera system and work in real-time. So, three main modules:

  • Track people with / without face masks.
  • System should be real-time and work for multiple cameras simultaneously.
  • Generate alerts for people not wearing a mask against their unique ids.

How does it work?

  • Camera Connection

    Our system connects to a network of cameras via a computer and processes all the video streams.

  • Faces & ID’s

    Dedicated APIs for lenders to provide real-time lending decisioning recommendations to clients

  • Alert

    The system marks and generates alerts against people not wearing masks.

Solution

We proposed to make a C++ framework based on TensorFlow. A C-API to handle multiple streams so the system remains state-of-the-art, and easy to scale up/down.

We focused on

  • Gathering face detection and mask datasets.
  • Make an alert system API to send information of people without masks.
  • Train a face detector and a mask classifier.

Challenge

Like with any project, there are always challenges. With this system we had to think of solutions against challenges like

  • No high quality mask datasets available for accurate prediction in the wild.
  • Running the system in real time on 4k Videos.
  • Create a C++ Application and Library that would be deployed on any system.
  • Small face detection in CCTV footage
Flow from an input image through person detection, tracking, cropping, face detection, and mask classification

Applications

  • Offices

  • Schools

  • Libraries

  • Hospitals

  • Restaurants

  • Industry

  • Hazardous Areas

Client Testimonial

RevolveAI was amazing to work with and we very highly recommend people looking for a top-talent in machine learning to work with RevolveAI! They built a great rapport with us from day one and we felt so comfortable with the whole team. Their experience and skillset, along with a laid back attitude and passion for AI technology, were immensely valuable throughout the process. They were very knowledgeable and gave us some great ideas and contributed so much to our project. In a nutshell I would say: Hire RevolveAI, you won’t be disappointed.