As a group, we've always wanted to create an app that makes use of facial recogntion. We conceptualised a fun way of using this technology by building an app that would search a database of roasts and toasts and match them to the user. Using the power of facial recognition in Amazon's Rekognition, HumbleMe can match the user's face to one in our database in order to produce either a roast (a light-hearted insult) or a toast (a compliment).

To build it, we made an app using React and Node.js for the front end, used PostgreSQL for the database, and Python's PRAW package to scrape for insults and compliments on Reddit using its API.

Some of the challenges we faced including using Amazon's Rekognition facial recognition technology as this was completely new to us. As well, we had difficulty with creating the database that would hold both the insults/complements and the source photos we used to match with the user's.

We are proud of being able to build our app from start-to-finish in only one day. This is an app that we feel could legitimately be featured on an app store and used by lots of people.

HumbleMe will continue to be a project we would like to grow and expand upon.

Share this project: