Automated_Headcount

Using MS machine learning REST API's to not only count the number of people present at a meeting, but how emotionally engaged they are

Comment

Inspiration

One of us is a TA, and Fatemeh thinks it is very challenging to both conduct a class meeting and to keep track of how many students attend each session. Keeping track of attendance by asking people to sign a paper by the door doesn't sit right with a Computer Science student, so together we came up with a modern solution to a mundane task.

What it does

Automated Headcount is an Android app that allows students to be effortlessly marked as present in a class or event. No extra effort has to be made by students or the meeting leaders, except the initial signing up onto the app.

How we built it

We used Android Studios to create a simple layout for our mobile app. In the initial installation, the app redirects the user to the Student or Admin side in order to set up the correct account. Students share their student id, and we make sure the student id is validated, and the Admins set up their email and personal password. To prevent non-Admins from setting an Admin account, we set implemented an authentication mechanism that's only known by the Admins. The attendance of the classroom is taken by Admin taking photos of the classroom. We used Microsoft's Cognitive Service Face Detect and Identify APIs to analyze the photo for 1) Who is present at the meeting and 2) Who is emotionally engaged with the contents of the meeting.

Challenges we ran into

We were very excited and hopeful to utilize the local area Bluetooth of mobile devices to take attendance! We ran into problems in the stage of making the Admin mobile into a surveillance device for all nearby devices who have installed the app and agreed to share their information via Bluetooth. We couldn't confirm the mobiles devices go offline as soon as their Bluetooth is switched off.

Accomplishments that we're proud of

Debugging Android Studio combined with Face (Azure Cognitive Services). There were many peculiar quirks to iron out.

What we learned

We learned a lot of technical skills such as debugging Android studios and debugging portions of code related to API calls. In terms of project life cycle, we got first hand experience with making tough decisions about changing the direction of project and being OK with the new goals in a short amount of time.

What's next for Automated_Headcount

We look forward to implementing an iOS app as well as generating a useful meeting summary and broadcasting it to all the attendees via their app. Although we have many more ideas and directions we'd like to take this app, the stated two are the most clear goals!

Built With

Share this project:

Updates