Electronic Theses and Dissertations

Date of Award

1-1-2022

Document Type

Thesis

Degree Name

M.S. in Engineering Science

Department

Engineering Science

First Advisor

Yixin Chen

Second Advisor

Dawn Wilkins

Third Advisor

Timothy Holston

Relational Format

dissertation/thesis

Abstract

The current facial recognition algorithms struggle with accuracy on real world cases. Haar cascade based algorithms are fast, but require fine tuning per image in order to achieve the best results. When tasked with images where there are multiple faces at different locations, the current algorithms seem to underreport the number of faces. This study attempts to produce a more accurate classifier through the use of taking the maximum result of multiple Haar cascade classifiers with differing parameters. To do this, a web image scraper was written to gather real world images from Google images and Flickr. These images were analyzed using the OpenCV library utilizing multiple Haar cascade classifiers and the maximum of these classifiers was taken. The result is a more accurate classifier, as most of the inaccuracies were due to undercounting, rather than overcounting.

Concentration/Emphasis

Computer Science

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.