Inspiration

We were inspired by the variety of apps that use revolutionary technologies such as AI and CV to make the world spin. We also saw that people want to find just the right song to help them destress, which can often be difficult.

What it does

The app takes a picture of the user's face, analyzes it for emotional expression, and then interprets this, keeping the user's past preferences in mind using a preference index calculated by multiplying the prevalence of the emotion by the database value. It then selects a song from a certain genre to fit the user's personality and disposition. After every song, the user is asked to rate their experience, and this data is put into the database. As a result, as the user uses the app more and more, the recommendations become more and more finely tuned to the user's tastes, as they relate to how they feel. A rating of five stars will cause a certain genre to be played more often in response to a certain emotion, while a rating of one will stifle that genre until it gets a chance to be played again. These factors combine with a fluid Android XML layout to provide an immersive experience for the user.

How I built it

We used Firebase backend for storing user data, and Google Cloud APIs to implement the computer vision. We also used Android camera libraries for the frontend.

Challenges I ran into

Handling GitHub's merge issues, as associated with Gradle build problems. Also API integration issues with Android Studio and Gradle.

Accomplishments that I'm proud of

Building a functional app using our own machine learning algorithm.

What We learned

We learned a lot about GitHub merge error handling, advanced machine learning algorithms

What's next for StressOpress

We plan to expand from music recommendations to recommendations for TV shows, places to go, and even movies.

Built With

Share this project:
×

Updates