MIME: Machine Intelligence for Motion Exegesis

MIME is a collaboration between faculty in the Department of Theater and Performance Studies at Stanford University and AI Modeling & Inference developers at the Stanford Libraries, with the goal of studying recorded theatrical performances through the lens of pose estimation and motion tracking. More information is available at our GitHub repo.

This portion of the MIME interface supports focused searching for specific sampled poses and hand gestures that are derived from "in the wild" performance recordings, predefined archetypes, and user input, as well as "distant viewing" analytics that can be applied across all or a portion of the entire corpus of performances. You may select the "Timelines" tab at left to access the other half of the site, which explores timeline and clustering analyses of poses, hands, actions, and (optionally) faces within individual performances.