Automated Detection and Classification of Multichannel Lungs Signals using EMD

Pulmonary disorders affect human life widely and are the third leading cause of death worldwide. Statistics show an annual mortality rate of 1.26 and 0.62 deaths per million person-years in men and women, respectively, due to Pneumothorax, 9 million due-to Bronchitis. Pneumonia is the single most significant infectious cause of death in children worldwide. Pneumonia killed 808 694 children under the age of 5 in 2017. The diagnostic measures in this area are mostly done using Spirometry or Imaging Techniques (X-Ray Photography), which is hazardous for the Human body. Our core objective behind this research is to propose a method that accurately classifies Pulmonary disorders. In this research, Lungs Sound (LS) signals were acquired using Stethoscope and Microphone. Samples from both healthy and diseased individuals are collected and analyzed. The acquired signals are first preprocessed and segmented using empirical mode decomposition (EMD). Mean of Mel-frequency cepstral coefficients (MFCCs) and Gammatone cepstral coefficients (GTCC) are the key feature to distinguish each class efficiently. Testing and training are performed using k-nearest neighbors (KNN), which accurately classified signals into healthy and diseased. This Multi-channel method is 99.5% accurate in disease detection and classification. This approach for Pulmonary disease detection is more precise, accurate, and safe.