Automatic detection of sleep apnea Disease from ppg signal using Optimization echniques

Abstract

Quite recently, the rate of people suffering from respiratory diseases has greatly increased no matter the age Generally, while breathing, the respiratory sounds are formed in the large airways and the combination of air velocity and instability tends to produce vibrations in the airway pipes. Later, these vibrations pass through the lung tissue and thoracic wall which an be used to observe sound using a stethoscope, micro phone or any other sensors. Typically, human respiratory signal is categorised into three types such as Sleep Apnea (SA), Motion artifacts and standard respiration. Among the three types, Sleep Apnea (SA) occurs during sleep because of monotonous syndrome at upper airway tubes. These abnormalities result in frequent awakening and oxygen de-saturation. When affected by Sleep Apnea (SA), the person pauses his breath by holding the air up to a minute and then releases the air. The breath suppression happens throughout their sleep. Illnesses like high blood pressure, heart attack, obesity and diabetes may cause the person to suffer from Sleep Apnea (SA). In order to detect Sleep Apnea (SA), several methods have been implemented. Here, four novel methods are proposed to detect Sleep Apnea (SA) precisely. Initially, Adaptive Neuro-Fuzzy Inference System (ANFIS) is used for classification of respiratory signals using cepstral coefficients. It is a novel and efficient method which classify normal and abnormal breathing sounds. As known, computerized methods of respiratory sound analysis will provide a quantitative basis for abnormal breathing sound recognition. In this work, the breathing sound signals are analysed using Mel Frequency Cepstral Coefficients (MFCC) and classified by Adaptive Neuro-Fuzzy Inference System Classifier newline