Take5

• 

  Logo

• 

  Render of the face landmark system

Inspiration

Our generation, ourselves included, spend copious amounts of time in front of our computers. It’s not a good habit to have, but simply put, building the self-control to wrench oneself away from the screen is quite boring!

What it does

Take5 is an innovative app that records the user’s movements when browsing, and identifies instances of subpar browsing behaviour - leaning in too far, straining your eyes, staying at your computer too long - and corrects the user through a variety of possible punishments, ranging from playing embarrassing music to triggering fork bombs. All the while, our AI bot will chastise users whenever they misbehave, further teaching users to build good browsing habits. No WiFi? No problem! Take5 can run without the AI bot functionality, and all of the other functions of Take5 can run without being connected to the internet.

How we built it

Challenges we ran into

1. Devpost deleted our first markdown submission 🙁
2. Struggled with a variety of mouse modules, including pynput and mouse
3. Accessing the wrong camera - MacBooks can access iPhone cameras
4. Debugging global state storage systems was a pain
5. Coming up with an idea

Accomplishments that we're proud of

1. Creating an algorithm to calculate the user’s distance from the screen based on just webcam footage.
2. Creating an AI bot using Groq and AWS in about an hour

What we learned

1. Various new technologies, including the specific mediapipe solutions, Groq, and PyAutoGUI.
2. The art of pivoting to better ideas, even if that means starting over.

What's next for Take5

Further implementation of a chatbot which gives advice about healthy habits while studying. More variety of punishments and punishable behaviours.

Built With

• amazon-web-services
• groq
• mediapipe
• opencv
• python