COVID RAKSHAK APP
Web based all in one app that can be used for Face Mask Detection, Social Distancing Detector, tracking COVID positive patients using centralized cloud data.
Created on 20th September 2020
•
COVID RAKSHAK APP
Web based all in one app that can be used for Face Mask Detection, Social Distancing Detector, tracking COVID positive patients using centralized cloud data.
The problem COVID RAKSHAK APP solves
The COVID Rakshak App is a viable solution for efficient functioning for the post COVID world, where we can monitor the public using live feed from the CCTV Cameras whether the citizens are following social distancing norms.
The app also has the feature to store data in the database and visualize data using python libraries like seaborn and matplotlib. One prominent feature about this app is that it runs on centralized cloud storage.
We can also visualize the data of COVID patients, Social Distancing Violators through our local web based site.
This app can also perform face recognition of the patients who are COVID positive or those people who are at risk.
It also provides valuable statistics of the surveillance so that we can automate most of the surveillance
without actually being physically present in the location and putting themselves at risk.
To provide a solution for minimizing workforce to be deployed for maintaining law and order in the COVID world and automate the process to provide a risk-free environment for ourselves.
This app uses the existing CCTV camera network to automate the surveillance.
It performs face recognition and alerts the user if he/she is not wearing a mask.
If the user is not wearing a mask, the computer system will speak up "Please wear a mask!".
It also checks for COVID positive/at-risk patients and stores the time, date and location of the person found in cloud storage.
We can use this app as a tool to make a hassle-free experience to keep a track of social distance violators, patients roaming freely or people who are not wearing mask - all in one app, the COVID Rakshak App. The app is very lightweight and is very easy to setup on computer system and it can be up and running with very little requirements. This app will ensure the safest way possible for surveillance in the post COVID world. Do watch the demonstration video for fully understanding the exact working of our app.
Challenges we ran into
The face detection process is an essential step as it detects and locates human faces in images and videos. The face capture process transforms analogue information (a face) into a set of digital information (data) based on the person's facial features. The face match process verifies if two faces belong to the same person. The problem we ran into was detecting faces with the wrong name. The bug was fixed when we trained the model using sufficient lighting conditions.
The app also has the feature to store data in the database and visualize it using python libraries like seaborn and matplotlib, all with a click of a button! The challenge we faced was making the data centralized. We solved this issue using one stop cloud storage, so as if one administrator edits the database, it is automatically updated for the other administrator.
The face recognition module was consuming more than 15 minutes to gets the model trained. And hence, it was practically very hard to implement in our app. We overcame this dilemma by using pre-trained models, which gave us a pretty good accuracy of above 90. Therefore, The face recognition module detects faces using pre-trained model stored in .yml format. Despite all the obstacles and hurdles, at last we were able to complete our project and were able to demonstrate a fully fledged working app.