How often have you asked someone how they’re feeling and gotten a quick “good” in response, knowing that they weren’t okay? How often have you been the one saying “good”?

This app was inspired by a desire to help make it simple for people to understand what is going on with themselves and with those that they care about, and to foster more positive interactions between people and the internet as well as people and other people.

What it does

Happy is a mental health Google Chrome app that wants to help you become more aware of, and sensitive to, your emotions and the emotions of others.

Happy works by using machine learning and computer vision to compute the likelihood that you are feeling emotions such as joy, sorrow, anger, or surprise, based on your facial expression.

How I built it

I built Happy using the Open Computer Vision 3 (Open CV 3) machine learning library for face detection and the Google Cloud Platform Machine Learning Vision API for sentiment detection. JavaScript was used as a client side scripting language, and HTML 5 and CSS were used to create the user interface.

Challenges I ran into

One challenge I ran into was capturing a snapshot from the video stream. It was extremely difficult to find a resource to address this challenge, as there was no documentation, forum discussion, or working open source code available. It was actually during a tangential conversation about photography that we figured out a solution - the Canvas API.

Accomplishments that I'm proud of

I’m proud of finishing a meaningful and ambitious project over the course of just 36 hours using technologies that were new to me!

What I learned

I learned how to build and debug a Google Chrome App. I also learned how to request and capture video through the computer’s webcam, take a snapshot of that video, and then send that snapshot to the GCP Cloud Vision API.

What's next for Happy

The next steps are to add more features and, if Happy gets some traction, to try and sell it on the Chrome Web Store!

Possible additional features include:

  • Analytics for sentiment over time so that people can begin to recognize patterns in their mood, as well as what is causing those patterns.
  • Alerts for parents when their children are unhappy for significant periods of time.
  • Alerts when you are visiting a site that frequently makes you unhappy. Suggestions for content similar to content that has made you happy in the past.

Built With

Share this project: