Automated mammographic mass shape classification using wavelets and neural networks

The objective of the thesis is to demonstrate the utility of wavelet transforms with artificial neural networks for the classification of mammographic mass shapes. A fully automated mammographic classification system has been developed to distinctly classify mass shapes as either round, which typically indicates the absence of breast cancer, or irregular, which typically indicates the presence of cancer. First, a wavelet transform was applied to the radial distance measure (RDM) of the mass shapes to obtain multiscale decomposition. The discrete wavelet transform and wavelet packet decompositions were investigated. The second step was the computation of scalar-energy features from the wavelet coefficients. Thirdly, a neural network classifier was used to classify the shapes as either round or irregular. A two-layer neural network with a backpropagation algorithm was trained on the wavelet-based feature vectors extracted from the RDMs of the mammographic mass shapes. A pilot study was conducted to investigate the effects of mother wavelet selection on the performance of the neural network classifier. As a final step, the performance of the automated classification system was studied using receiver operating characteristic (ROC) analysis. (Abstract shortened by UMI.).

Name thesis:degree_name

Master of Science (MS)

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Electrical and Computer Engineering

Grantor dc:publisher

University of Nevada, Las Vegas

Year

2000

Author dc:creator

• Shanmugam, Nithya

Contributors dc:contributor

• Lori Bruce

Statement dc:rights

• IN COPYRIGHT. For more information about this rights statement, please visit http://rightsstatements.org/vocab/InC/1.0/

Language dc:language

English

Identifier

https://oasis.library.unlv.edu/rtds/1157

OAI identifier oai:identifier

oai:oasis.library.unlv.edu:rtds-2156