Virtual Headcam (VHC) is a device designed for facial performance capture. VHC equips with two motorized mirrors (mirrors mounted on motors) with rotational axes aligned perpendicularly and a fixed camera pointing at one of the mirrors. The two motorized mirrors can redirect the optical path of the mounted camera. With a motor controller, the stationary camera can virtually pan and tilt in a certain angular range in a tracked space volume.
VHC can be integrated into a motion capture space where the location of actors and actresses can be accurately tracked in high refresh rate. VHC can quickly adjust the mirror angles, zoom, and focus, according to the motion tracking data centered the subject in the camera frame. A system of multiple VHCs installed in a motion tracking space can potentially replace the traditional head mounted camera system.
This research project was one of my work projects at USC ICT, and it was presented at SIGGRAPH 2015 poster session.
For more detailed of this project, please visit the project page on USC ICT Graphics Lab website.