For the first build, I was focused on getting clmtrackr working with p5.js. This demo features realtime face tracking with some effects drawn over the face.
Version 2 builds on the first by incorporating multiple different types of emoji to render face elements in real time.
Version 3 achieves a mask-like effect by tracing the outline of the face in addition to emojis that track facial features.
Version 4 incorporates real-time emotion tracking. The relative strength detected for each emotion determines the size at which it's displayed.
Version 5 builds on the last by using data from the emotion tracker to render the dominant emotion over various face components in real time.
Version 6 creates a circular frame effect of reactive emoji's surrounding the face
Version 7 features a halo of emoji around the head and a stream of recent emotions that pours from the mouth when open.
In version 8, I've built on the code to generate emojis that float around the screen when the mouth opens, based on the current dominant emotion.
Version 9 features a large emoji that covers the face and updates with the current dominant emotion.