This work introduces a novel deep learning framework for identifying and classifying automated lung disease from CT scan images. The proposed technique employs a ResNet-50 backbone with dual attention to improve the normal, benign, and malignant lung condition feature discrimination. The methods of attention-guided feature extraction, thorough preprocessing, and interpretable visualization via Grad-CAM contribute to the proposed work’s superior performance. Experimental results demonstrate the value of the proposed approach for supporting clinical diagnosis. Keywords: Deep Learning, Lung Disease Detection, Attention Mechanism, Medical Image Analysis, CT Scan Classification