Prostate cancer is a leading cause of cancer death for men in the United States. Fortunately, the survival rate for early diagnosed patients... Show moreProstate cancer is a leading cause of cancer death for men in the United States. Fortunately, the survival rate for early diagnosed patients is relatively high. Accurate prostate cancer localization with noninvasive imaging using MRI can be used to guide biopsy, radiotherapy, and surgery as well as to monitor disease progression. However, in general, a single type of MRI is not sufficient for reliable tumor localization. As an alternative, multispectral MRI, i.e., the use of multiple MRI-derived datasets, has emerged as a promising noninvasive imaging technique for the localization of prostate cancer; however almost all studies are with human readers. There is a significant inter and intra-observer variability for human readers, and it is substantially difficult for humans to analyze the large dataset of multispectral MRI. To solve these problems, this thesis presents various novel supervised and semi- supervised (interactive) segmentation techniques. Initially, we develop a supervised segmentation method by combining conditional random fields (CRF) and support vector machines (SVM) with a cost-sensitive framework, and show that proposed method further improves classical and cost-sensitive SVM results by incorporating spatial information. Next, we propose an extension of popular semi-supervised seg- mentation method, namely random walker (RW) algorithm, with automated seed initialization for multispectral MRI images. We also present an automated shape and boundary based segmentation approach for prostate segmentation from T2-weighted MRI. Proposed method is based on a banded geocuts algorithm that utilizes bound- ary and shape information to yield prostate segmentation. Finally, we develop a novel method that has the ability to design classifiers obtained from one imaging protocol and/or MRI device to be used on a dataset from another protocol and/or imaging device. In order to evaluate the performance of the proposed methods, we utilize multispectral MRI datasets acquired from 21 biopsy-confirmed cancer patients. Our results show that multispectral MRI helps to increase the accuracy of prostate cancer xiv localization when compared to single MR images; and that using advanced proposed methods for prostate cancer localization performs better than available methods in the literature. PH.D in Electrical Engineering, May 2013 Show less
