Pulmonary Nodule Detection and Analysis using X Ray Image Processing

Abstract

The World Health Organization World Health Organization (WHO) states that newline newlinelung cancer is a prominent cause of cancer-related deaths worldwide. The employ- newlinement of image processing methods and Computer-Aided Diagnosis Computer-Aided newline newlineDiagnosis (CAD) systems have improved the identification of lung nodules. Chest newline newlineradiographs can be used to detect several chest ailments, including TB, adenocarci- newlinenoma, squamous cell carcinoma, big cell carcinoma, and atelectasis. The accuracy newline newlineof lung segmentation plays a crucial role in the creation of a CAD system. This newlinestudy presents an unsupervised learning method for segmenting the lungs in chest newlineradiographs. The method utilizes a circular window and local thresholding. The newlinetechnique involves pre-processing, an initial estimation of the lung area, and the newlineremoval of any unwanted interference. The images are first resized to a resolution newlineof 1024x1024 and then improved using adaptive histogram equalization. The chest newlineradiographs are converted into binary images using the suggested technique. The newlinelungs are distinguished from the chest radiographs based on their geometric and newline newlinespatial properties. The last stage in image segmentation involves employing mor- newlinephological operations. Local thresholding, excluding unnecessary body parts, filling newline newlinein gaps, and filtering areas based on their characteristics all contribute to initial newlineestimations of the lung field. Morphological methods are employed to eliminate newlineextraneous noise. The performance of the proposed technique is evaluated using newlinea publicly available Japanese Society of Radiological Technology (JSRT) dataset, newlinewhich includes 247 chest x-rays, after removing the bone shadow. The suggested newlinemethod s efficacy is assessed by comparing its findings with those of Active Shape newline newlineModel (ASM)-based lung segmentation using several performance metrics, includ- newlineing F-score, overlap, accuracy rate, sensitivity, specificity, and precision rates. All newline newlinethe parameters for the suggested technique exceed 90%. The research imply that newlinethe proposed metho