13 February 2012 Face recognition via a projective compressive sensing system
Author Affiliations +
A projective compressive sensing system for face recognition is presented. The Fisherfaces method was utilized for classification of the face images. The system uses a digital micromirror device to project measurement vectors onto the scene and a single photodetector to collect the backscattered illumination. Experimentally, the system accuracy was 95.5% using only 32 measurements per image; this performance matches the simulation results. The total number of image pixels was 5,736 (84 × 64) resulting in a compression factor of 168 over a conventional imaging system.
© (2012) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Brant M. Kaylor, Brant M. Kaylor, Charlie J. Keith, Charlie J. Keith, Peter A. Roos, Peter A. Roos, Randy R. Reibel, Randy R. Reibel, } "Face recognition via a projective compressive sensing system", Proc. SPIE 8254, Emerging Digital Micromirror Device Based Systems and Applications IV, 82540G (13 February 2012); doi: 10.1117/12.909056; https://doi.org/10.1117/12.909056

Back to Top