Publication

Abstract : This study offers a promising technique for the categorization of respiratory sounds, which may significantly impact healthcare outcomes. This system’s deployment is possible in various healthcare settings, including distant or rural places. The study stresses using the models on the Microcontroller, which allows for real-time categorization of respiratory sound prediction. The respiratory sounds in the study are divided into multiple categories using deep learning methods such as Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU) models. The models’ good classification accuracy was examined after they were trained on a sizable dataset of auscultation sounds. The CNN and GRU models obtained 95% and 93% accuracy, respectively.