Inspiration

According to reports, there are 200 different kinds of underlying mental disorders that cannot be easily detected out of which Depression, Anxiety Disorders and Dementia are the most common. Shockingly, 1 in 5 people in the US is affected with some sort of mental disorder where 25 million people are solely affected by only depression! An even shocking fact is that 7% of people who experience depression commit suicide 😥. And that's what we are trying to say that Mental disorders are big and Mental health matters...

After researching a lot we found an accurate identifier that will partially help in tackling this problem and it is Emotion. Emotions play an important role in our life, Apart from just expressing ourselves, they do a lot more. According to Harvard, Emotions are the best representation and revelation of one’s mental health. Emotions are also the easiest way to find about someone’s mental conditions. Using the inter-relation between emotions and mental health we developed an app and we called it EmotionSwitch

What it does

EmotionSwitch is a machine-learning-powered tool that uses powerful and complex algorithms to keep track of your emotions and mood swings to detect a probable underlying mental disorder before it's too late. All of this is done just from your webcam video feed..

All you need to do is visit the website and start a session after allowing your webcam to be captured then you can carry on with your other tasks while the web app keeps a real-time track of your emotions, you can also see your emotions varying on a realtime graph that is visible on the web app. After a suitable time, you can end the session when you feel like (an average good session is 15 minutes). After ending the session, the app will generate a detailed report with a lot of content including a potential mental disorder detection and some visual content including a graph and a pie chart. The app also provides a downloadable transcript which you can share with your therapist later on.

image

How we built it

Challenges we ran into

Accomplishments that we're proud of

We are very proud of the following things:

  1. Completing such a complex app within 36 hours
  2. Very proud of our modelling approach, that's what makes us different from existing solutions
  3. Integrating a video-based ml model without WebSockets directly into the flask
  4. Having a nice UI
  5. Making a real-time graph

What we learned

We learned a heck of a lot of stuff during the hackathon and we're extremely thankful for that, some of the major stuff we learned:

  1. We played with and learned a lot about flask and compiling a full-stack app
  2. We mastered frontend skills and using frameworks to speed up the development process
  3. Integrating a video-based machine learning model
  4. Hosting complex and heavy full-stack apps online.
  5. Team coordination skills

What's next for EmotionSwitch

We had a lot planned for EmotionSwitch but we couldn't finish all of it in 36 hours. So we will cover that parts here:

  1. Improving the interface and design of the app
  2. Adding Medical records to the system and securing them.
  3. Having it more personalized and saving all previous sessions.
  4. Test prototype with a psychologist/psychiatrist

We really had fun making this app. All of the code was written 100% by us

Built With

Share this project:

Updates