Date of Award

Spring 2010

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Electrical & Computer Engineering

Program/Concentration

Electrical Engineering

Committee Director

Vijayan K. Asari

Committee Member

Frederic D. McKenzie

Committee Member

Jiang Li

Call Number for Print

Special Collections LD4331.E55 A426 2010

Abstract

A new face recognition algorithm using a synthetic discriminant function based shifted phase-encoded fringe-adjusted joint transform correlation (SDF-SPFJTC) technique is proposed. The dark region in an input image is enhanced by using a nonlinear technique named ratio enhancement in gaussian neighborhood (REIGN). Histogram equalization and Gaussian smoothing are then performed to the enhanced face images and the synthetic discriminant function (SDF) image before they are subjected to the joint transform correlation process. The two distinct correlation peaks produced on extreme ends of the SPFJTC plane signifies the recognition of a potential target. A post processing step utilizes the peak-to-clutter ratio (PCR), the magnitude of the maximum correlation peak and the second maximum correlation peak, along with their location in the x-y plane to determine if the test face belongs to a known perpetrator. Performance of the proposed face recognition technique is verified using the Yale facial expression database with varying illumination and facial expressions. Computer simulations show efficient and successful recognition accuracy in varying environmental conditions. Research work is still in progress to obtain recognition results to support pose-invariant face recognition and to obtain an automatic threshold determination technique unique to each individual.

Rights

In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/ This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).

DOI

10.25777/d1qk-0k11

Download

Share

COinS