9 June 2011 Lossless divide and conquer for time-order sorting of N facial poses
Author Affiliations +
Abstract
Law enforcement agencies need a quick decision in May 1 2010 NY Time Square for quickly spot wanted car bomb individuals in a crowd. This goal requires real-time smart firmware and a smart search algorithm to know how to order the faces in a geometric way. We demonstrate such a sorting problem of N facial poses time ordering is like the TSP of N cities, NP-complete, having no exact deterministic solution. Here we demonstrated a heuristic working solution to answer the ONR grand challenge called Empire 2010. How could the N boxes of faces detected and cut by the efficient COST parallel color-hue-algorithm, without time mark, at a single CPU (or even with the time marks but collected videos from multiple vintage points) to determine automatically who speaks what, where, and when? There must be a cross platform sensory association from our different sensors.
© (2011) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Ming Kai Hsu, Ming Kai Hsu, Harold Szu, Harold Szu, } "Lossless divide and conquer for time-order sorting of N facial poses", Proc. SPIE 8058, Independent Component Analyses, Wavelets, Neural Networks, Biosystems, and Nanoengineering IX, 80580W (9 June 2011); doi: 10.1117/12.887551; https://doi.org/10.1117/12.887551
PROCEEDINGS
7 PAGES


SHARE
RELATED CONTENT

Technology survey on video face tracking
Proceedings of SPIE (March 02 2014)
A multi-camera system for real-time pose estimation
Proceedings of SPIE (April 29 2007)
Recording The Evidence
Proceedings of SPIE (February 19 1973)

Back to Top