Selin Ugur, Mohammad Safaei
Lung cancer is one of the deadliest forms ofcancer, and early detection can make a life-saving difference.This paper introduces a practical and innovative method fordetecting lung tumors in CT scans using a combination ofgenetic algorithms and convolutional neural networks (CNNs).By leveraging genetic algorithms to fine-tune the CNN’s hyper-parameters, our approach achieves higher accuracy and reli-ability in identifying tumor regions. We trained and validatedthe model on publicly available datasets and included visualexplanations, such as saliency maps, to show how the AI makesits decisions. The paper also compares our results to existingmethods and highlights the potential of this technique in real-world medical applications. With its ability to provide clear andinterpretable results, this method could help clinicians detectlung tumors more efficiently and with greater confidence.