Lung Cancer Detection And Classification

Lung cancer remains one of the most critical health concerns worldwide, accounting for a major share of cancer-related deaths each year. Early and accurate detection is essential for improving survival rates, yet traditional diagnostic methods such as X-rays, CT scans, and biopsies often face limitations, including high costs, invasiveness, and dependence on expert interpretation. These challenges highlight the need for automated, reliable, and scalable approaches to support timely diagnosis. This project introduces an AI-driven Lung Cancer Detection System that combines machine learning and deep learning techniques to classify lung images into benign, malignant, and normal categories. Using a dataset of 9,000 CT images from Kaggle, the system applies preprocessing and segmentation methods to enhance image quality, followed by feature extraction using the Histogram of Oriented Gradients (HOG). Classical models such as Random Forests and Decision Trees are compared with deep learning architectures like Convolutional Neural Networks (CNNs), DenseNet, and ResNet. Results indicate that deep learning models, particularly CNN and ResNet, outperform traditional methods in accuracy and robustness. By offering real-time predictions through a web- based interface, the system reduces manual workload and supports radiologists in faster, more consistent, and effective decision-making.