Off-campus South Dakota State University users: To download campus access theses, please use the following link to log into our proxy server with your South Dakota State University ID and password.
Non-South Dakota State University users: Please talk to your librarian about requesting this thesis through interlibrary loan.
Document Type
Thesis - University Access Only
Award Date
2015
Degree Name
Master of Science (MS)
Department / School
Electrical Engineering and Computer Science
First Advisor
Sung Y. Shin
Abstract
Breast cancer classification technique divides breast cancer into two categories, benign tumors and malignant tumors. The main purpose of breast cancer classification is to classify abnormalities into benign or malignant classes and thus help physicians with further analysis by minimizing the possible errors that can be done because of fatigued or inexperienced physician. Irregular shape is one of the most frequently appearing feature for the malignant masses which can be used to identify breast tumor as benign or malignant. The main objective of this thesis is to introduce new improved shape features that can be used to classify mammogram images into normal (benign) or cancerous (malignant) classes with better classification accuracy compared to that of traditional ones. In this paper, we propose five new shape features variance and range based on measures of variability, convexity index, circularity range ratio and irregularity ratio to effectively classify breast cancer images. Support Vector Machine (SVM) is then used as a machine learning tool to train data first and then classify breast cancer images. The main contribution of this paper is to improve the classification accuracy method over existing traditional shape features. A total of 500 images obtained from ix DDSM (Digital Database for Screening Mammography (DDSM, University of South Florida) with a total of 300 training and 200 test cases. Firstly, binary object images were extracted from actual breast mammograms using Support Vector Machine. In order to evaluate the improved performance of the proposed shape features, convexity, circularity and a modified global shape feature of compactness was used. The result obtained shows that the proposed shape features can improve measure of performances such as MCC, specificity, sensitivity and accuracy and can be a promising tool to provide preliminary decision support information to physicians for further diagnosis.
Library of Congress Subject Headings
Breast -- Radiography
Breast -- Imaging
Breast -- Cancer--Diagnosis
Image processing -- Digital techniques
Image analysis
Radiography
Medical-- Digital techniques
Description
Includes bibliographical references (pages 41-48)
Format
application/pdf
Number of Pages
86
Publisher
South Dakota State University
Recommended Citation
GC, Sailesh, "Breast Cancer Classification of Mammographic Masses Using Improved Shape Based Features" (2015). Electronic Theses and Dissertations. 1848.
https://openprairie.sdstate.edu/etd/1848
Rights Statement
