I’m experimenting with using my phone’s camera to do hand pose recognition for an upcoming piece at NYCEMF 2023 in June. In this video I’m triggering scenes in Ableton’s session view by using a simple classification model in Wekinator, which interprets data sent from a Node-for-Max & Mediapipe set of example patches made by lysdexic-audio and made available on Github.
Hopefully as I get comfortable I’ll be doing more interesting things with this.