Deep Learning for Early Lung Cancer Detection from CT Scans: A Data Science Bowl Approach

Lung cancer remains a significant global health challenge, with early detection being crucial for improving patient outcomes. This paper presents the development of computer-aided diagnosis classification algorithms to detect cancerous lesions in lung CT scans. Leveraging advancements in deep learning and transfer learning, I focus on extracting features from pretrained neural networks to improve classification accuracy. My methodology involves preprocessing CT scan images, employing convolutional neural networks and gradient boosting machines to classify nodules, and evaluating model performance on a Kaggle-hosted dataset from the Data Science Bowl 2017. I discuss challenges such as data normalization and feature extraction from three-dimensional images and outline future directions for refining this approach. Overall, this study contributes to ongoing efforts to harness artificial intelligence for enhancing early lung cancer detection and underscores the potential of deep learning techniques in medical imaging analysis.