Pattern Recognition in Lung Disease Classification: Leveraging Convolutional Neural Networks that Employs Local Phase Quantization Features

Lung diseases are wide group of diseases related to the respiratory system, and is ranging from acute infections to chronic disorders with multifactorial causes. Recognition of the lung disease is very crucial in the proper diagnosis, forecasting through prognosis, and remedies. Hence proposed lung disease classification with feature extraction technique. The proposed work carried out feature extraction with Gray-Level Co-occurrence Matrix (GLCM) and Local Phase Quantization (LPQ) combined with Convolutional Neural Network(CNN). The goal is to create a model that reduces diagnostic errors, ensuring more precise early detection and classification of pulmonary disorders. This initiative facilitates the advancement of pulmonary medicine by promoting accurate radiologic imaging, while future directions emphasize integrating medical workflows to enable personalized and timely diagnosis and treatment. Additionally, the ResNet model demonstrates superior performance in hernia detection, achieving a test accuracy of 89.43% for GLCM based feature extraction and Pneumonia detection with 79.20% test accuracy for LPQ based feature extraction.