Leveraging Deep Learning Techniques for Early Lung Cancer Detection from Computed Tomography Images

Lung cancer represents a critical health condition that can result in fatality if not identified during its initial stages. Early detection of lung cancer allows for effective treatment options. Symptoms associated with lung cancer generally manifest in the human organism during its terminal phase; however, advancements in technology and computer-assisted systems can identify it at an earlier juncture. The process of diagnosing lung cancer necessitates enhancement, and there exists a demand for a cost-effective and user-centric system that utilizes cutting-edge data science methodologies. The deep convolutional neural network has been designed to extract significant features that enhance the detection and diagnosis of lung cancer cells with greater efficiency and accuracy. The methodology proposed in this project has yielded promising and commendable outcomes regarding accuracy, precision, recall, F-score, and specificity metrics in lung cancer identification. Databases with varying views of the computed tomography (CT) scanning system-axial, coronal, and sagittal-have been established, and the suggested method is built with pulmonary pictures obtained from persons with and without cancer. The proposed research demonstrated an overall accuracy of 95% to 98%. The proposed CNN algorithm can potentially save lives by improving the accuracy of medical diagnostics and aiding in the early diagnosis of lung cancer.