Microsoft’s Emoji8 app uses Machine Learning to test your facial expression skillz

Microsoft has uploaded a small and fun sample app to the Microsoft Store which demonstrates their Windows Machine Learning API, which does local on-device (vs server-based) machine learning to, in this case, recognize facial expressions.

The Emoji8 sample UWP app uses Windows Machine Learning to evaluate your facial expressions while you imitate a random selection of emojis.

This app showcases how to leverage the low-latency, scalable, and flexible performance of Windows ML. Frames are locally assessed as soon as they are snapped from the camera to allow real-time machine learning evaluation and game screen updates.

The app is quite fun and in the end, you can even share an animated gif of your best faces to Twitter.

The source code is also available at GitHub here. In addition, the ML model used for this local evaluation, FER+ Emotion Recognition, can be found at Github here.

Find the app in the Store here.

Price: Free

Source, Thanks Venkat for the tip.