Classification of asthma severity levels using pulmonary acoustic signals based on wheeze sound

Asthma is a common pulmonary disease that affects approximately 300 million people worldwide each year. At present, there are no cures for asthma, and therefore, the cornerstone of asthma management is to achieve optimal asthma control. Every well-controlled patient should have an asthma action plan (AAP). It involves a paper written asthma diary to monitor and record the daily symptoms of asthma based on the Peak Expiratory Flow Rate (PEFR) measured using a Peak Flow Meter (PFM). The PFM however not been implemented due to some limitations such as it restrictions to be used by severe asthma patients and also asthmatic babies does hinders the effectiveness of the AAP itself. Other than PFM, stethoscope is also used in diagnosing asthma severity. Even though it is reliable, only medical personnel can interpret the respiratory pathology from the pulmonary acoustic signals. Therefore, this thesis discussed on the development of a computerized decision support system (CDSS) to classify asthma severity using pulmonary acoustics signals based on wheeze sound. This system is designed based on the relationship between PEFR and also wheeze sound that have been reported by some of the previous researcher. In order to develop the CDSS tool, two frameworks were proposed and the most effective framework was used to implement for the development of the CDSS for classification of asthma severity levels. Both of the suggested frameworks were proposed based on the standard frameworks for image/signal processing which involved preprocessing, feature extraction and classification process. However for framework two, there is an additional process of wheeze segmentation method. The proposed wheeze segmentation method adapted Short Time Energy (STE) and Fuzzy Inference System (FIS). The data collected were pre-processed in accordance to the Computerized Respiratory Sound Analysis (CORSA) guidelines.