FPGA implementation of eye tracker video processing

Current wearable video video-based eye tracking technology has considerably strict limitations in size and processing speed. Hardware used to process video must be lightweight, compact and somewhat modular. The video should be processed as it is being collected in real time and accurately enough to generate useful, reliable data. The utilization of general purposegeneral-purpose logic processors makes the implementation of basic pupil detection and eye tracking algorithms simple and compact, but it introduces accuracy issues due to the necessary simplicity of the computational methods used to detect and track the pupil. Large, complicated hardware implementations are more accurate, but they are quickly outdated and unwieldy. Application Application-specific programmable logic devices solve these issues in part since they allow users to synthesize a fast hardware device that realizes complex and robust algorithms. The devices are also capable of being updated or redesigned quickly and easily, as often as needed. When applied to eye tracking, this could act to improve on the accuracy issues and maintain functionality within a small, self-contained physical device. A basic video processing device implementation was attempted within severe functional constraints for thi