Feature Selection and Segmentation for Posterior fossa tumors

Over the last few years segmentation of brain tissues and abnormalities has been an important area in medical imaging. Several morphological imaging techniques have been used to solve this problem. Many fractal based features have been used to analyze brain tumors. These fractal based methods have improved segmentation of brain tumor. This book provides features selection techniques based on statistical model for selecting best features from subset of features and use it for segmentation. Two statistical methods based on Kullback Leibler Divergence and Bayesian models have been analyzed along with segmentation techniques such as SOM and EM. The analysis should help professionals who are working for computer analysis diagnosis of brain tumors.

Autorentext

Shaheen Ahmed received M.S.and Phd in Electrical & Computer Engineering from Wright State University,OH & University of Memphis, TN, USA in 2006 and 2011 respectively.Her research interest includes image processingfocusing on medical images, application of pattern recognition for analysis of MRI, and statistical methodologies.