Increase MediaPipe's detection model complexity

[lddm]

Hi @cgtinker, this tool is fantastic. Thanks for sharing it publicly 👏

I'm having issues with the MediaPipe pose detection accuracy. The landmarks jump all around the place while tracking the pose of a moving person. I can obtain more accurate pose detection by using Mediapipe on my own, outside of Blender. I'm using mediapipe==0.10.2, and setting the model_complexity=4. But then, I don't know how to load the detection into BlendArMocap to

Did you ever consider allowing for model_complexity values greater than 1? Would it make sense to enable loading the landmarks detected offline by using a .csv file or something similar?