Dissertation/Thesis Abstract

Refining Bounding-Box Regression for Object Localization
by Dickerson, Naomi Lynn, M.S., Portland State University, 2017, 55; 10602491
Abstract (Summary)

For the last several years, convolutional neural network (CNN) based ob- ject detection systems have used a regression technique to predict improved object bounding boxes based on an initial proposal using low-level image features extracted from the CNN. In spite of its prevalence, there is little critical analysis of bounding-box regression or in-depth performance evaluation. This thesis surveys an array of techniques and parameter settings in order to further optimize bounding-box regression and provide guidance for its implementation. I refute a claim regarding training procedure, and demonstrate the effectiveness of using principal component analysis to handle unwieldy numbers of features produced by very deep CNNs.

Indexing (document details)
Advisor: Mitchell, Melanie
Commitee: Liu, Feng, Massey, Barton
School: Portland State University
Department: Computer Science
School Location: United States -- Oregon
Source: MAI 57/01M(E), Masters Abstracts International
Source Type: DISSERTATION
Subjects: Computer science
Keywords: Computer vision, Convolutional neural network, Localization, Machine learning, Object detection
Publication Number: 10602491
ISBN: 978-0-355-43935-9
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest