Inspiration
we love emoji
What it does
We analyze user's input facial expression and return an close matched emoji
How we built it
We crawled data from CMU Cohn-kanade database, and build the model on Clarifai. Then we wrote the front end that allow user to upload or take real-time video stream to get feedback from Clarifai api.
Challenges we ran into
Building the proper training model for facial expression
Accomplishments that we're proud of
We improved traditional facial recognize model with weighted algorithm.
What we learned
Clarifai Api, Computer vision, python, javascript
What's next for Emoji Matcher
Integral more facial expressions with higher accuracy.
Log in or sign up for Devpost to join the conversation.