OVERVIEW

Our project goal was to simplify the emoji selection process so that users do not have to search through all emojis anymore to find a suitable one. Therefore, we created an app that uses a neural network to detect facial expressions of users and transforms them into an emoji. We built and trained a neural network for six different emotions using a combination of multiple available emotion data sets and images that we scraped from the google image search to increase accuracy. We also tried to include more than six emojis but it was not possible to achieve a satisfactory accuracy with publicly available data sets.

We then included our model in an android app and implemented the UI to enable users to take a picture and generate an emoji with it.