Face Detection

Challenge

 

The team was looking for making their sign up solutions more authenticated. They needed the profile picture of their users to have a human face in it rather than having an object. This solution needs to be a robust and a reusable solution for facial detections but also localization of a face in a given image.

Solution

 

Using a modular approach an API was developed which could be implemented out of the box. Powering the API was a machine learning model trained to detect the faces in an image and also localise it. Its solution goes deeper and identifies a number of faces and also localises each of the faces landmarks eyes nose ears and mouth separately.

Technologies Used

Face Detection Uses

  1. Sign up solution 
    • Checking if the user has uploaded a face or an object.
  2. Population counters
    • Count the number of distinct faces in a given frame.
  3. Automated doors
    • Automatic door switches when a human comes near.
  1. Motion detection
    • Detecting the movement of the landmarks as they move
  2. Surveillance
    • Detecting if a human has appeared in an under surveillance area
  3. Authentication processes
    • Any authentication process requiring a human face to be assured.