Computer Aided Detection of Lung Nodules in Ct Images Using Swarm Intelligence

Abstract

ABSTRACT newlineLung cancer is one of the most common forms of cancer resulting in over a million deaths newlineper year worldwide. The most critical aspect related to the occurrence of cancer is the use newlineof tobacco. Persistent of cough, shortness in breathing, pain in the chest, and bronchitis are newlinesome of the symptoms exist when the cancer is in advanced stage. It is found that if the newlinedisease is detected earlier, about one-third of cancers become preventable, another one-third newlinebecome potentially curable. newlinePulmonary nodules are small tissue in the lung and most of them are benign. In medical newlineimaging, detection of lung nodule is one of the most challenging tasks. Examination of newlinelung nodule is performed with various imaging modalities such as radiography, computed newlinetomography, magnetic resonance imaging and positron emission tomography. Among them newlinecomputed tomography is most frequently used imaging modality due to low cost, quality newlineand robustness. Due to non-pathological structures the radiologist finds difficult and timeconsuming newlineto distinguish some nodules in computed tomography. newlineTo attain a more reliable and accurate detection, computer assisted approach can be newlinedeveloped to assist interpretation of the medical images. The computer aided detection is newlinea potential method to accomplish a range of quantitative tasks such as early cancer, disease newlinedetection and analysis of disease progression. It has more advantages in terms of speed newlineand accuracy in detection of pulmonary nodules in computed tomography images and thus newlinereduction in miss rate. newlineThe main objective of this research is to develop an automatic computer aided detection newlinesystems to differentiate pulmonary solid, part solid and non solid nodules from non newlinenodules from computed tomography images. The developed system comprises of segmentation newlineof lung nodules from segmented and reconstructed lung parenchyma, extraction of newlinefeatures from the segmented nodules and their classifications using particle swarm based newlineneural network classifier. newlineSegmentation of lung parenchyma is employed usi