A System for Automated Reduciton of MPEG-1 Encoded Facial Videos

Not Available We propose to design and build an analog VLSI smart-pixel wavefront-sensor chip that will drive a deformable membrane mirror to correct rapidly changing wavefront distortions. On-chip detectors will sense the error signal in local wavefront tilts using a Hartmann-sensor within a feedback loop. The error signal is then used to drive the membrane mirror to the phase conjugate of the incoming distorted field. The chip wil use a massivley parallel analgo computation algorithm to determine the error signal present during closed loop operation, when the turbulence is probed with light from a glint or a with a guide star. Our approach integrates the Hartman sensor, membrane mirror, and analog computation onto a single die substrate allowing for a high-speed, ultra-light weight, compact, low-power adaptive optic system. Our chip-driven membrane-mirror approach promises up to 8p radians of stroke, 10us response time and true-piston modulation. In Phase I, we will simulate, fabricate and test separately a proof-of-concept 10x10 smart-pixel chip. We will interface a membrane mirror modulator to the chip to demonstrate an optical wavefront correction system capable of 100 Khz operation. In Phase II we will scale the device to a 256x256-element array and demonstrate wavefront correction for atmospheric turbulence and retinal imaging. We plan to market this device for military countermeasures, seeker systems, and low-cost commercial retinal imagers.