Using Pose and Action Estimation to Analyze Theater Performances

Watch a Stanford HAI research seminar exploring how machine learning and computer vision techniques analyze directorial styles in theater performances through pose and action recognition. Learn how scholars combine distant and close viewing approaches to study video recordings of theatrical productions, identifying patterns in choreography and staging that reveal a director's unique artistic choices. Discover an innovative methodology that bridges performing arts and computer science, offering new perspectives on theatrical analysis and the systematic study of gestural patterns in performance. Recorded at Stanford University on January 22, 2025, this hour-long presentation demonstrates the potential of computational tools to enhance our understanding of theatrical artistry and directorial decision-making.