Model based segmentation and bioinspired approaches for the diagnosis of pulmonary disorders from computed tomography images

Abstract

The advancements in the field of information technology and newlinemedicine have paved the way for the use of computers in diagnosis. It has newlinealso become important to reduce the mortality rate by early detection of newlinediseases. Computer aided diagnostic (CAD) systems are being developed to newlinediagnose diseases from different medical imaging modalities such as chest newlineradiographs, computed tomography (CT) images and magnetic resonance newlineimaging (MRI) scans. Interpretation of these images presents challenges to newlinethe expert as the patterns of the affected regions may look similar for one or newlinemore diseases. CAD systems have been developed to diagnose the disorders newlinepresent in lungs, brain, liver and spinal cord. CAD systems can be used by newlinephysicians as a second opinion in decision making and treatment planning. newlineIn this research work, three CAD systems to diagnose pulmonary newlinedisorders and a segmentation approach have been proposed. In the first work newlinecarried out, a CAD system has been proposed to diagnose drug sensitive newlinetuberculosis with cavities, consolidations, nodules and military pattern newlinemanifestations from CT images. In the second work, a CAD system to newlinediagnose bronchitis has been developed and in the third work, a CAD system newlineto diagnose pulmonary hamartoma has been proposed. The fourth work is a newlinemodel based segmentation approach which focuses on complete segmentation newlineof lung parenchyma, thereby increasing the accuracy of segmentation newline newline