Multi-body tracking with Mediapipe/other libraries

Luke Buzard's icon

I'm currently working on an installation where one major part of it is using the positions and movements of an "audience" to drive some sort of synth(s) . Currently, I'm using the Mediapipe templates from https://cycling74.com/forums/n4m-facemesh-handpose-google-mediapipe. However, I can't seem to figure out how to get Max to differentiate between different people.

Does anyone know how to--or if it's even possible with Mediapipe--to separate the streams of data from different bodies? Or is there another piece of software that might work better for this application? Outside of Max, I have very little programming experience, so I've been having a hard time understanding anything when I look it up.

Would appreciate any sort of help from anyone knowledgable! Thanks!