AgentSTAR: Agentic Shape Tracking and Reconstruction from Monocular Videos
This work presents a method for shape reconstruction and tracking from video via agentic analysis-by-synthesis that infers a structured 3D object model, including its geometry and kinematic structure, and uses this model to optimise object track estimates over time.
Kirill Mazur, Nikita Karaev, Matthew Chang et al.
· 1 citation