Automatic detection of lung nodules in computed tomography images using U-Net

Lung cancer is one of the leading causes of cancer-related deaths methods for lung nodules in computed tomography (CT) images rely on manual interpretation by radiologist, which can be time-consuming and prone to human error. This paper presents BreathSafe.AI a deep learning system for the automatic detection and segmentation of lung nodules in CT images using an enhanced U-Net architecture combined with dense network techniques. Our model is trained on the LUNA16 dataset, utilizing advanced image preprocessing and segmentation methods to optimize nodule detection. This system achieves a diagnostic accuracy of over 90%, significantly improving detection speed and consistency compared to existing methods. The results highlight the system’s potential to enhance lung cancer screening by reducing diagnosis time and variability, making it valuable tool for clinical use. Our approach demonstrates superior performance compared to state-of -art techniques, offering a scalable and efficient solution for early detection of lung cancer.