Lung Cancer Classification Using Deep Learning: A Comprehensive Approach with Modified Convolutional Neural Networks

This paper presents a strong deep learning technique to identify lung cancer through computed tomography (CT) imagery, using the IQ-OTH/NCCD lung cancer set of data. We have come up with a convolutional neural network (CNN) that is to classify CT scans in normal, benign and malignant categories. The CNN architecture consists of multiple convolutional and pooling layers, dropout mechanisms as a regularizer and fully connected layers for feature extraction. We perform comprehensive data pre-processing which involves image resizing in order to improve model performance. The proposed model having achieved excellent accuracy during testing periods. These results reveal how deep learning could revolutionize the way we diagnose and classify lung cancer accurately with CT scans.T By arguing for the application of deep learning in clinical work to increase accuracy of diagnosis, the research shows just how significant use of artificial intelligence has become in the analysis of images taken during medical procedures. This proposed CNN based approach achieved a test accuracy of 99.55%, precision of 100%, recall of 99% and a f1-score of 98%. This approach might help radiologists make diagnoses sooner and plan treatments better, so patients do well because doctors can treat them quickly and accurately.