A Deep Learning Model for Lung Cancer Detection using CNN

Deep learning has transformed medical research. It provides powerful, disease-specific functions that are now widely used in healthcare. Among a number of the most important fields is the early detection of lung cancer where the deep learning method, especially convolutional neural networks (CNN’s), has profoundly changed the way of diagnosing of diseases. These recent advances significantly increased the accuracy and speed of detecting lung cancer nodules in CT images. Here we are interested in distinguishing cancer versus non-cancerous lung nodules, using CT scan images in our work and deep learning for such task. To improve prediction accuracy and generalization, we adopted an ensemble approach. It combines several CNN designs and allows deeper analysis by leveraging the strengths of different models. We used publicly available dataset, which have well-annotated CT scan images to develop deep learning model. The annotated CT grayscale images served as input for the model. These images provided references that helped the system learn intricate patterns and features related to lung cancer. To ensure reliable evaluation, we carefully divided the dataset into training, validation, and test sets. This allowed us to systematically test the performance of the model. The three CNN used in our ensemble model, called lung net, were created with different numbers of layers, kernel sizes, and pooling in order to extract feature representations of distinct characteristics.