Wrapper Based Feature Selection for CT Image

Diagnostic imaging is invaluable. Magnetic Resonance Imaging (MRI), digital mammography, Computed Tomography (CT), and others ensure effective noninvasive mapping of a subject's anatomy, and increased normal and diseased anatomy knowledge for medical research in addition to being a critical component in diagnosis and treatment. In this work various feature selection algorithms are investigated and a Swarm Intelligence Algorithm based on Bacterial Foraging is proposed. Features are extracted using wavelet and Gray-Level Co-occurrence Matrix (GLCM). The obtained features are fused using Median Absolute Deviation (MAD) after normalization and the feature selection techniques investigated. Results obtained show the improved performance of Bacterial Foraging based feature selection for different classifiers.