Graphical User Interface for the CVIP MATLAB Toolbox with Application to Syrinx Detection in Veterinary Thermographic Images

Abstract

The Computer Vision and Image Processing software tool (CVIPtools), developed under the direction of Dr. Umbaugh at Southern Illinois University Edwardsville (SIUe), is a powerful application used in both research and teaching [1]. CVIPtools is composed of C language functions to perform imaging operations that are divided into libraries based on function categories; a Common Object Module (COM) was added to each function and then included into a dynamically linked library to be used with the Windows operating system [1]. The CVIPtools C functions were implemented in MATLAB as the CVIP MATLAB toolbox (CVIP Toolbox) containing a set of functions for image processing in MATLAB [2]. This project uses CVIPtools version 5.7e as a reference, version 2.5 of the CVIP Toolbox and the MATLAB Graphical User Interface Development Environment (GUIDE) for development. Here, the development and testing of the Graphical User Interface (GUI) for the functions available in the CVIP Toolbox are described.

The development includes the design of the MATLAB figures for each image operation or group of operations, and the code required to capture the parameters given by the user in the GUI and call the functions in the CVIP Toolbox. This thesis also includes some bugs found in the CVIP Toolbox functions, the debugging that was done to fix it, and the creation of new functions to improve the functionality of the GUI [3].

This project was developed following the menus and image operations available in CVIPtools version 5.7e. The first part of the project is focused on methodology for providing Windows behavior to the CVIP Toolbox GUI according to the options available in MATLAB, followed by the design of the MATLAB figures to perform image analysis operations such as the geometry, edge/line detection, segmentation, transforms, feature extraction and pattern classification. A GUI was developed to perform compression related operations such as the image preprocessing and postprocessing, and the implementation of the block truncation coding algorithm that belongs to the lossy compression methods. The GUI contains operations included in the Utilities menu of CVIPtools such as the arithmetic and logic, compare, convert, create, enhance, filter, size and stats image operations. All the image operations and functions were tested in several binary, grayscale and color images with power of 2 and non-power of 2 sizes. Results of image operations in the GUI were compared with the results obtained from the same operations using CVIPtools, and a special focus was given to the functions that perform feature extraction and pattern classification. Finally, one application example is provided of the use of the CVIP Toolbox GUI in the classification of thermographic images of canines for Syrinx detection.