Comparative Analysis of Deep Learning Models for Lung Cancer Detection: DenseNet121, AlexNet, and VGG16

This research investigates the submission of CNNs for the accurate organization of lung cancer from medical images. Leveraging three distinct architecturesDenseNetl21, Alex Net, and VGG16-this study evaluates their performance on a dataset of CT scan images. The dataset, comprising 364 training images and 190 testing images, was preprocessed using data augmentation techniques to enhance model generalization. DenseNet121, Alex Net, and VGG16 were meticulously constructed and trained, each incorporating a series of convolutional, pooling, and dense layers tailored to optimize feature extraction and classification accuracy. The models existed assessed based on their accuracy, precision, recall, F1-score, and AUC metrics. Remarkably, DenseNet121 achieved the highest accuracy, outperforming Alex Net and VGG16, which also demonstrated commendable performance. The results underscore the efficiency of CNNs in the automated detection of lung cancer, suggesting a substantial potential for enhancing diagnostic procedures in clinical settings. Each model's architecture and training processes were critically analyzed to understand their respective strengths and limitations. The conclusions underwrite to the growing body of information on AI applications in medical imaging, if a substance for future advancements in the field. In conclusion, the integration of DL techniques for lung cancer recognition presents a promising path for improving early diagnosis and treatment products. This study highlights the superior performance of DenseNet121 while recognizing the valuable insights gained from Alex Net and VGG16, paving the way for further research and development in medical AI applications.