Meerkat

Our inspiration

School is a place where kids can learn and have fun, and time is of the essence. Ensuring the safe arrival of every student is a must, especially in an environment such as school. Additionally, attendance can waste precious time that could instead be used to learn, study, or participate in fun activities. We wanted to change that.

Our application

Meerkat is an app that uses facial recognition to take attendance as students walk into the class instead of having the teacher waste precious time for a the lesson everyday doing it, or forgetting to do it at all. The attendance system records student's names on the screen as they walk into class, and it can check in multiple students at one time. The app can also account for students signing out and signing in during class, to go to the bathroom for example. In order to preserve battery, as soon as the connection between the camera and the app expires, the application shuts down shuts down. We built it using android studio for the app, python and OpenCV for facial recognition, and web sockets to communicate between the camera and app through WiFi.

Challanges

The first AI trained was not accurate at predict faces because backgrounds for data sets were very different, so we had to make new data sets with the same background. We had difficulties implementing camera functionality on android studio at first, so we changed to WiFi. The biggest accomplishment we are proud of is overcoming all the obstacles and ending up finishing the project successfully. One major thing we learned was how to create an app in android studio. Another major thing we learned was that instead of spending a lot of time to research how to get a task done one way, it might be easier and more efficient to take a step back and use a different approach. The next goals for Meerkat are new camera integration, faster training algorithms, and bigger data sets for each student, increasing accuracy. New data sets could be taken from security camera footage.