SBIR Phase I: Breaking the Barriers to the Commercialization of Super-Resolution Video Enhancement Algorithms

Machine Visionaries, Grand Forks ND

Investigators

Abstract

This Small Business Innovation Research (SBIR) Phase I project will investigate the two primary barriers to the commercialization of super-resolution video enhancement, namely, that (1) the motion compensation-based video enhancement algorithms are extremely computationally-intensive, and (2) the ability to extract additional visual information from digital video is highly dependent on image sequence content. Provided that people, objects, and entire scenes move in subpixel increments between video frames or fields, this motion can be exploited by temporally integrating neighboring pictures to increase spatial resolution and to actually see details where there were once blocky pixels. The proposed research focuses on determining the feasibility of significantly increasing the processing speed of the associated numerical algorithms on desktop and notebook personal computers running Microsoft Windows or Linux, as well as determining the feasibility of automatically detecting the video frames/fields which contribute additional information to a specific region-of-interest selected for enhancement. A number of commercial applications in forensic image analysis for law enforcement and security end-users can be developed around this technology. Super resolution video enhancement is of particular interest to government agencies, such as the Department of Defense and the Department of Justice for extracting additional details from reconnaissance and surveillance video, as well as to the private sector for forensic image analysis in law enforcement, surveillance, and security. By successfully tackling the proposed technical objectives, the capability will be developed to embed super-resolution enhancement video technology into commercial nonlinear video editing software packages such as Adobe Premiere.