Pose Animator is a free and open source implementation showcasing how you can animated an SVG avatar using either webcam driven face and bone data, or from analyzing a static image. On the GitHub repository, Pose Animator is described accordingly:
Pose Animator takes a 2D vector illustration and animates its containing curves in real-time based on the recognition result from PoseNet and FaceMesh. It borrows the idea of skeleton-based animation from computer graphics and applies it to vector characters.
This is not an officially supported Google product.
The GitHub page also includes two demos. The first one showcases using a camera to control an avatar, while the second one showcases how to accomplish the same thing with a static image. Please note there a currently a couple bugs that need to be addressed if you build this from source yourself, with the resolution details available here. I also walk through the process of building and running Pose Animator in the video below.