Di-Mask: A Real-time Social Distance Estimation (in inches) and Mask Detection Pipeline, which uses Machine Learning to help prevent the spread of COVID-19. #preventcovid #socialdistancing
The COVID cases are increasing day by day and it is very important for the government to keep a track of cases. This motivated me to build an app that detects masks and also checks if social distancing is being followed correctly or not.
The project is capable of identifying if people are wearing masks or not. It shows the status as N if they are not wearing the mask and shows Y if they are wearing the mask. It also tells the approximate distance in inches between 2 people which helps us keep a track of social distancing being followed correctly. I have linked the video based on live streaming of camera demonstrating the mask status and distance in inches between two people.
The application was build using the Viola-Jones algorithm for face detection in Python, Flask, GCP i.e Google Cloud Storage, Google Vision API, Google AutoML, Google API engine, and few other technologies like HTML, CSS3.
It was difficult to translate the distance captured through the camera to realtime distance in inches
I am happy that my application is useful not only in health care but also in scenarios where we have social gatherings or corporate offices, schools, colleges for mask detection and check if social distancing is being followed correctly.
Di-Mask is capable of detecting the masks for crowded situations. The application uses Viola-Jones and Google Cloud Platform. Viola jones is used to detect the faces and these faces are then passed to GCP for mask detection done through Vision API. I am happy that I was able to integrate it correctly.
Test it for a large number of users and different lighting scenarios
Google Cloud Storage, Google Vision API, Google AutoML, Google API engine
Python Flask
Machine Learning
UI : HTML5, CSS3