Enhanced Respiratory Tract Auscultation Audio Signal Classification Technique Employing LSTM and RNN

The complications and fatalities caused by respiratory tract infections are a major public health issue worldwide. Detecting the wide variety of respiratory disorders, from pneumonia to LRTI, URTI, and COPD, requires cutting-edge diagnostic tools. Thanks to advances in AI technology and the improvement of medical diagnostic tools, early detection of such diseases is now possible. In this article we offer a state-of-the-art LSTM-based RNN model that has been meticulously trained on more than 1800 lung audio recordings. We've used data augmentation methods to make sure our dataset is as complete and accurate as possible. In addition, we have combined MFCC and genetic algorithms in our feature extraction method to select the most relevant features for our model. Our findings are conclusive; the suggested MFCC-GA-LSTM model achieved an impressive 98.65% accuracy and an F1 score of 99.16, proving its superiority over prior approaches to classification. The potential for early detection and prevention of these potentially fatal illnesses has been greatly improved because to this novel approach we have developed for diagnosing infections of the respiratory system.