Collaborators:
Petros Faloutsos (EECS, Lassonde Scool of Engineering, York), Costas Sideris, (Dept. of Computer Science, UCLA) CS, Erik Dutson (UCLA Dept. of Surgery)
Decription:
We are investigating methods for accurately tracking from video the spatial motion of all objects that take part in FLS Training Tasks. By exploiting the geometric and photometric invariants common to standard FLS training environments, we can robustly and accurately track the instruments and all manipulated objects directly from the FLS station's video stream, with no modifications to the standard FLS training box, camera or instruments.
