MATLAB-based implementation of a lightweight deep learning model (RDLINet) for classifying lung sounds into respiratory disease categories using Mel spectrograms. Achieved 84% accuracy.