Inspiration
I've tutored students in math and physics for many years and very early on, it quickly became evident that these students were incredibly bright, but they just never asked for clarification in class. Often in tutoring sessions, I am stuck wondering whether a student is confused or not and oftentimes it hinders the progression of the class. I can only wonder how it must be when it is not just a one on one interaction.
What it does
Automatically takes attendance through facial recognition software using only one image. For deployment purposes, the image could be just a yearbook photo which nearly everyone takes.
Scans the emotions and facial expressions of the students in the classroom and tells the teacher when the majority of the students are confused.
How I built it
For the frontend, a MVVM structure was created with UIKit in swift. The backend was built with python and flask. The facial recognition model came from openCV and the emotion recognition model was the VGG19 model trained on the FER-13 dataset.
Challenges we ran into
The first main challenge was using flask for the backend because I had never done it before but it quickly became evident that it wasn't very difficult and was quite similar to HTTP requests in JavaScript. The next main challenge was getting both AI algorithms to run in tandem with each other and produce a combined output that could eventually be extracted with the HTTP requests. The final hurdle was creating all the niche features like tracking all the total time the students are in the classroom and warning the teacher about each student that is skipping class.
Accomplishments that we're proud of
Building a fully functioning, ready to deploy app that can be used in public schools today in under 24 hours.
What's next for teach.ai
I plan on trying to take this company forward. This is not just some idea I had for only a hackathon, but rather, my vision is much greater. This tool would be extremely useful all accross the country and could cause extreme changes in how our future generations grow and develop their minds. I plan on building a few fully working prototypes and seeing if I can try and deploy them in the schools near my house. From there I can build a reputation and scale the company further.
Log in or sign up for Devpost to join the conversation.