In this thesis, we propose a task-based approach to parametric imaging with dynamic image sequences and apply the proposed method to an... Show moreIn this thesis, we propose a task-based approach to parametric imaging with dynamic image sequences and apply the proposed method to an example problem of prostate cancer segmentation with dynamic contrast enhanced Magnetic Resonance Imaging (DCE MRI). Traditionally, the time activity curve obtained from dynamic series of MR images is modeled without considering a specific task in order to obtain the kinetic parameters and to construct the parametric images. This mostly consists of estimating parameters based on minimizing the error between the model and measurement. Unfortunately, this method results in noisy images and performances of the task in hand e.g. tumor segmentation suffers. Therefore, we develop a new method for the estimation of kinetic parameters based on the maximization of tumor segmentation performance. The kinetic parameters are estimated with a weighted approach such that the performance of the particular task is maximized. The mathematical criterions used to quantify the performance are Fisher Ratio, Area Under the Curve (AUC) and the Dice Measure. The proposed method of parametric imaging is tested with the problem of prostate cancer localization with DCE MRI. We use Linear Discriminant Analysis (LDA) as a segmentation tool and use quantitative measures to compare segmentation results, such as Fisher Ratio, Dice Measure, Sensitivity, Specificity and Area Under the Curve (AUC). Our results show that the proposed method is able to improve the prostate cancer localization for certain patients. M.S. in Electrical Engineering, July 2011 Show less
Query
(-) mods_name_creator_namePart_mt:"Haleem, Muhammad Salman"
