Angel

Home page
Reminder to excercise every 30 minutes when you've been in front of computer screen for too long
Eye tracker detects distraction
Bad posture blurs your screen
Detects sleepiness and encourages better sleeping habits
Daily quiz for mental health
Daily quiz for physical health
Test Results
Now a chrome extension!

Inspiration

When working at home alone, we often fall into bad working habits. We fall asleep during zoom meetings, can't focus anymore for long durations, and don't ever get up from the computer. We no longer have someone to motivate or push us to do our best while locked up at home. We especially need to stay healthy during these times and build up our immune systems.

What it does

Angel is a chrome extension that keeps you accountable by watching out for you. There are two main features, the ai powered webcam and mental and physical health quiz. You first enter your work period hours in the popup, and press start. The AI-powered webcam then blurs your screen whenever you have bad posture, notifies you when it notices your gaze is distracted or you are looking at your phone for too long, tells you to go to sleep if you look tired, and reminds you to exercise if you've been sitting in front of the screen for too long. After your work period is over, the chrome extension prompts you to take a quiz. This quiz allows you to freely write about how you're feeling and your workday and provides a quote that directly addresses what you've written. you may also click on a body part that might be aching after your long hours, and the results may return some possible treatments. These features all contribute to work productivity, which leads to happiness and motivation.

How we built it

To analyze posture, we used PoseNet and created an algorithm that takes into account both the patterns that show on your face when you bend forward or how your head moves when your back curves. We also trained a model on detecting sleepiness using ml5js and used an eye-tracking library to detect patterns of distraction (ex, when the user looks at their phone for a long period of time). We also determined average movement of the user during sepcific times using PoseNet, so that we can apporpriately notify the user to excercise if they've been in front of the screen for too long. When the user inputs how they're feeling in the quiz, the bitext ml model predicts the most important words and emotions and we use those to find a suitable quote in our database using cosine similiarity. Finally, we used apimedic to find diagnosis for any parts of your body that might be hurting after a long day of work.

Challenges we ran into

We had a lot of issues with poor documentation in the webgazer library used to track distraction and the apimedic api. When transitioning our demo into a chrome extension, we came across a lot of permission issues as well.

What's next for Angel

We plan to make angel into either a separate camera monitoring device or a computer app so that it isn't limited to chrome. We also want to display or send a report every day to the user.

Prize Money

If we were to win prize money from "In Code We Trust", we would like to put it all marketing our chrome extension, which is in its first testing phase - Spade. We would also want to set apart a part of this to maybe make a working device with a camera for Angel that would work pretty much the same, but also link to your phone. As we don't know much about computer application development, we might have to use some of this money to hire developers that can put our features onto an application.

“ANGELS can turn our darkness to light, help our dreams to come true, and our worries take flight, so pass them your burdens, give them your cares, they’re waiting to help you, just trust they are there.”