Image-space coding of time and focus

Patrick Llull
Duke University

Given interesting features on temporal scales of microseconds and focal scales of centimeters, the data rates required to sample megapixel-scale, dynamic 3D scenes are tremendous.

We discuss sampling strategies utilizing code division multiple access of video to encode such scenes into 2-dimensional, video-rate measurements. We show how decompressive inference may be used to decode the multiplexed data to estimate the underlying dynamic scenes.


Back to Computational Photography and Intelligent Cameras