I mean this is it, this is the gimmick. Drawing in Air with the help of OpenCV for capturing and ORT for inference with the help of YOLO Hand Pose model trained for 800 epochs with the dataset given in the Ultralytics website.
ORT + training a YOLO hand pose model seems to be working
Having a hard time implementing AI reading Images for hand landmarks, initially tried MediaPipe bindings, but the project had been discontinued and had a bunch of errors from breaking changes of many dependencies, after trying to resolve them I ended up in a huge rabbit hole so I am trying other options now, read about ONNX and ORT, will try to implement something with it.
