LUNA-X: A Hybrid Deep Learning Framework for Automated Lung Nodule Detection, Disease Classification, and Three-Dimensional Pulmonary Visualization from CT Images
Madhu SivaSelvi S
LUNA-X is a hybrid deep learning framework that couples a ResNet backbone with a self-attention Vision Transformer (ViT) to automate pulmonary nodule detection and classification in CT scans. The system improves clinical decision support by combining detailed local texture analysis with global spatial dependencies, achieving a 95.60% Dice Similarity Coefficient on the LIDC-IDRI dataset for identifying malignancies, pneumonia, and fibrosis.