Connecting unsafe users to the nearest safe users during a natural disaster. When a user A marks himself as 'unsafe', other users in his vicinity are notified, if they mark themselves as safe, user A is notified of the nearest safe user along with the shortest path to the same. If too many users mark themselves as unsafe in a particular locality, all users in nearby localities are notified that a 'potential' disaster may be imminent.
Using state-of-the-art face comparisn models to match lost people with the people looking for them. People can upload pictures of people who are 'Lost' or 'Found' on our app. The faces are encoded to give feature vectors which are compared. If a 'Lost' face is found to be similar to a 'Found' face, both parties are notified and respective contact details of are provided.
Begin automatic crowdfunding whenever natural disasters are detected. When a major natural disaster is identified through reliable sources (GDACS), we begin automatic crowdfunding campaigns through our app where users can donate to specific disasters through paytm.
Machine Learning is employed in many places in our App. The risk generated is based upon predictions from Artificial Neural Networks. Computer Vision (the YOLO algorithm) is extensively used in damage detection and people counting. We also use Convolutional Neural Networks to encode facial features when comparing faces to help reconnect 'lost' people with their families.
GDACS is a cooperation framework between disaster managers worldwide to improve alerts, information exchange and coordination in the first phase after major sudden-onset disasters. We use web scraping to extract useful information from the GDACS website and this data is used in our Machine Learning predictions and in our automatic crowdfunding campaigns.