Inspiration
The Telus prompt really pushed us to look at the problem at a unique angle.
What it does
A machine learning algorithm detects the emotions of the user and send a JSON object of the detected mood and the percentage accuracy of the prediction. The p5.js sketch takes that JSON object and changes the amount of Perlin Noise of the drawing to attempt to abstract the emotion into some kind of representation.
How we built it
Challenges we ran into
Libraries, dependencies, and sometimes even storage space were some of the many issues that we had to face when working on theis project. Along with being a complete novice when it comes to Neural Networks, the project also involved processing which I personally had only heard of once before.
Accomplishments that we're proud of
After many hours, the neural network finally connected to the training dataset and was fully operational
What's next for Mirror Journal
Our next steps would be to create a platform that integrates all these individual moving parts together to flow as one.
Built With
- css
- html
- javascript
- keras
- p5.js
- python
- tensorflow
Log in or sign up for Devpost to join the conversation.