Prediction of Lung Cancer Using VGG19 Training Learning

Abstract

The critical need for accurate prediction of lung cancer, employing artificial intelligence on CT scan images to mitigate the high mortality rate associated with this disease. Deep learning, particularly CNNs, emerges as a powerful tool for achieving superior prediction accuracy compared to traditional machine learning methods. Leveraging a dataset comprising 3000 chest scan images across various types of lung cancer which includes adenocarcinoma, benign and squamous cell carcinoma the effectiveness of multiple machine learning algorithms is evaluated. It confirms CNN as the optimal choice for accurate prediction, with the implementation of VGG-19 further enhancing the assessment of lung cancer severity and precautionary measures. Performance analysis is obtained by using accuracy, precision, recall and loss metrics. For designing this application python software is used and result analysis is performed.