Optimizing A Deep Learning System For Lung Cancer Classification

Lung cancer is one of the leading causes of mortality worldwide. Early and accurate detection of lung lesions is essential to improve clinical outcomes. In this study, a convolutional neural network (CNN) model was developed to classify lung computed tomography (CT) images. The current study processed the dataset “IQ-OTH/NCCD_lung_cancer_dataset”. A strategy including data augmentation techniques was designed to optimize the generalization of the model. When evaluating the model on the test set, an accuracy of 0.9819, a precision of 0.9575, a recall of 0.9742 and an F1-Score of 0.9653 were obtained, indicating outstanding performance in image classification. These metrics were derived from the confusion matrix, which demonstrated minimal misclassification, particularly between benign and normal categories, while achieving perfect identification of malignant cases. These results suggest that the proposed model is a promising tool to support the diagnosis of lung lesions in CT images.