Deep Learning-Based Classification Methods for Prediction of Lung Cancer Images

Cancer, a disease with a devastating impact on lives, is often brought about by a confluence of hereditary disorders and other aberrant changes. Cancer is a sickness were body cells multiply too quickly. The brain and lymph nodes are among the organs where lung cancer may spread after starting in the lungs. Lung cancer changes depending on a number of medical factors and geographical places. The study presents a comparative study for lung cancer detection by CT (Computed Tomography) images, encompassing data collection from LIDC-IDRI (Lung Image Database Consortium and Image Database Resource Initiative). Through data collection, preprocessing, and splitting. Deep learning (DL) models including GoogleNet, AlexNet, CNN, and FocalNet are employed for classification. In comparison to the other models that were evaluated, FocalNet exhibited exceptional performance according to F1-score, accuracy, precision, and recall. It achieved outstanding scores of 99.81 %, 99.78%, 99.36%, and 99.56%, individually. The results of this study demonstrate that FocalNet is an effective method for computer-aided identification of lung cancers. Furthermore, this study highlights the potential value of FocalNet in clinical practice for the purpose of improving lung cancer detection and management. This study offers valuable insights into a development of efficient computational frameworks for medical image analysis, particularly in oncology diagnostics.