Date on Master's Thesis/Doctoral Dissertation


Document Type

Master's Thesis

Degree Name



Computer Engineering and Computer Science

Committee Chair

Elmaghraby, Adel S.


Diagnostic imaging; Imaging systems in medicine; Prostate--Cancer--Diagnosis


In this thesis, a computer aided diagnostic (CAD) framework for detecting prostate cancer in DWI data is proposed. The proposed CAD method consists of two frameworks that use nonnegative matrix factorization (NMF) to learn meaningful features from sets of high-dimensional data. The first technique, is a three dimensional (3D) level-set DWI prostate segmentation algorithm guided by a novel probabilistic speed function. This speed function is driven by the features learned by NMF from 3D appearance, shape, and spatial data. The second technique, is a probabilistic classifier that seeks to label a prostate segmented from DWI data as either alignat, contain cancer, or benign, containing no cancer. This approach uses a NMF-based feature fusion to create a feature space where data classes are clustered. In addition, using DWI data acquired at a wide range of b-values (i.e. magnetic field strengths) is investigated. Experimental analysis indicates that for both of these frameworks, using NMF producing more accurate segmentation and classification results, respectively, and that combining the information from DWI data at several b-values can assist in detecting prostate cancer.