Diagnosis of Lung Cancer by Using Maximum Senstivity Neural Network

Abstract

Various methods are employed for the identification of lung cancer. Computed newlineTomography (CT) scans, X-rays scans, Magnetic Resonance Imaging (MRI) scans, newlineand Positron Emission Tomography (PET) scans are few of the techniques used for newlinesame. Interestingly, CT scan images are favored for their efficacy in detecting small newlinelung nodules. Despite the availability of numerous models for lung cancer detection, newlinethere remains a constant need for improvement in terms of accuracy. This research newlineendeavors to address this gap by focusing on the development of a novel model newlinenamed the Maximum Sensitivity Neural Network (MSNN), with effective newlinepreprocessing and segmentation methods. Grayscale lung CT images are utilized as newlineinput for the MSNN model, and after rigorous training and experimental verification, newlineit outperforms existing deep-learning models, achieving an impressive accuracy rate newlineof over 96% and a sensitivity rate of 94%. The incorporation of sensitivity maps newlineserves to verify the accurate classification of cancerous and non-cancerous images. newlineFurthermore, geometric features are computed for the identified nodules to validate newlinesoftware outcomes. Therefore, this research makes a substantial contribution to the newlinefield of precise lung cancer detection, which has the promise of improving medic newline