Abstract
Lung cancer detection using Convolutional Neural Networks (CNNs) represents a transformative approach in medical imaging, utilizing deep learning to identify cancerous images in lung images. Our project aims to develop and implement a CNN-based system to detect lung cancer from computed tomography (CT) scans. The process begins with preprocessing the CT images to standardize and highlight key features, followed by training a multi-layer CNN designed to recognize complex visual images. For this project, the key technologies used include Python for scripting, Tensor Flow as the deep learning framework, and CPU acceleration to expedite model training. The CNN architecture consists of layers for convolution, activation functions, pooling, and dropout to reduce over fitting. The model is trained on a sizable dataset of labeled lung images, comprising a diverse mix of normal and cancerous cases, providing a solid foundation for robust learning. The final results indicate a high accuracy of 95%, signifying a highly effective model with low rates of false positives and false negatives. This level of accuracy demonstrates that the CNN-based approach can substantially support radiologists in early lung cancer detection, which could lead to improved patient outcomes through earlier diagnosis and intervention. The success of this system underscores the potential of deep learning to transform medical diagnostics and assist healthcare professionals in making more informed clinical decisions.
Keywords: Computed Tomography, Convolutional Neural Networks, Deep Learning, Lung Cancer Detection, Model Accuracy.