Enhancing Lung Carcinoma Screening: Utilizing 2D-UNet Models for Precise Detection

The goal is to revolutionize early detection methods for lung cancer, which is a major cause of global cancer deaths. It addresses the challenge radiologists face in analyzing numerous CT (computed tomography) scans, leading to fatigue and reduced accuracy. A cutting-edge solution is introduced to tackle this: a sophisticated 2D-UNet model within a user-friendly web application. The model isolates cancerous areas on reference images by employing precise cropping techniques, streamlining the detection process. Training the model with datasets from reputable lung cancer centers ensures reliability. Extensive experimentation with advanced techniques and hyperparameters fine-tunes the model's performance, which is evaluated through metrics like the dice coefficient and confusion matrix. Ultimately, the project aims to boost early diagnosis rates and improve patient outcomes by revolutionizing lung cancer detection methods.