Neural Network Models for Diagnosing Lung Diseases
Annotatsiya
This work provides a detailed description of the stages involved in the process of classifying lung sounds based on artificial intelligence for the diagnosis of various respiratory diseases: bronchiectasis, bronchiolitis, chronic obstructive pulmonary disease (COPD), pneumonia, upper respiratory tract infections, and healthy states. The stages of sound classification include: data preprocessing, data postprocessing, data splitting, deep learning model development, and model evaluation. The tasks of each stage are analyzed based on the optimization of lung sounds for organizing deep learning through the development of a convolutional neural network model. Methods and algorithms for improving the reliability of the developed neural network model are proposed. The suggested convolutional neural network architecture effectively analyzes patterns based on training and test datasets, leading to more accurate classification of respiratory diseases based on lung sounds. Future directions and improvements in the classification of lung sounds using artificial intelligence for diagnosing lung diseases are proposed.