Dissertation/Thesis Abstract

Development of a Computer Vision and Image Processing Toolbox for Matlab
by Gorantla, Lakshmi Anjana Devi, M.S., Southern Illinois University at Edwardsville, 2018, 71; 10844337
Abstract (Summary)

Image Processing and Computer Vision (CVIP) applications can be developed and analyzed using the CVIPtools software developed at Southern Illinois University Edwardsville in the CVIP Laboratory under the guidance of Dr. Scott E Umbaugh. The CVIPtools software has been created with the code in the C/C++/C# programming languages. Due to the popularity in engineering applications for Matlab use it was decided to port the CVIPtools libraries functions to Matlab M-files and create a CVIP Toolbox for Matlab.

This work consists of developing, testing, packaging, developing documentation for, and releasing the first version of the Matlab Computer Vision and Image Processing Toolbox. In this there are several steps involved which are described clearly in this research work. The primary aim of thesis work is to create a toolbox which is independent of any other toolboxes in Matlab. CVIPtools has over 200 functions which are written in C, but due to growing demand for Matlab we decided to make the functions available in Matlab. After the toolbox is created, the user can install it and can use the functions in the toolbox as Matlab inbuilt functions. This will make it easy for the user to understand and experiment with different CVIP algorithms.

Initially the toolbox was created writing wrapper functions for the programs written in C through the creation of MEX functions. But later due to problems during testing, it was determined [5] that it would be more suitable to write separate Matlab code, M-files for all the functions and create new toolbox.

The CVIP Toolbox for Matlab is an open source project and is independent of any other toolboxes. Thus, the user can install the toolbox and can use all the functions as Matlab inbuilt functions without the need to purchase any of the other Matlab toolboxes, which is required for other toolboxes of this type. There are 206 functions in this first version of toolbox which are the primary functions for CVIP applications. These functions are arranged according to categories so that it will be easy for the user to understand and search various functions.

The CVIP Toolbox is organized into several folders including CVIP Lab, which allows the user to create any algorithm with the help of functions available in the toolbox. The user can explore by using different functions in the toolbox and varying parameters experimentally to achieve desired results. The skeleton program for lab is in cviplab.m which has a sample function implemented so that the user can see how the sample is executed and can call other functions using the same method.

Indexing (document details)
Advisor: Umbaugh, Scott E.
Commitee: Engel, George, LeAnder, Robert
School: Southern Illinois University at Edwardsville
Department: Electrical Engineering
School Location: United States -- Illinois
Source: MAI 58/01M(E), Masters Abstracts International
Subjects: Education, Electrical engineering, Computer science
Publication Number: 10844337
ISBN: 9780438360327
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy