Browsing by Author "Kazemi, Mohsen"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    New prognostic index to detect the severity of asthma automatically using signal processing techniques of capnogram
    (Universiti Teknologi Malaysia, 2013) Kazemi, Mohsen
    Asthma is a chronic inflammatory disease of the bronchial tubes that happens approximately in 3% to 5% of all people in their life. Currently, capnography is a new method to monitor the asthmatic conditions, and unlike traditional methods, it is taken while the patient is breathing as comfortably as able. Previous studies concluded a significant correlation between the capnogram and the asthmatic patient. However, most of them are just manual studies conducted through the conventional time domain method based on the assumption that the capnogram is a stationary signal. However, manual analysis of capnogram is time-consuming and leads to erroneous results due to human factor such as tiredness. Therefore, a new prognostic index to automatic detection the severity of airway obstruction by processing the capnogram signal is presented in this research. First, in order to investigate the property of capnogram signal, the first and the second statistical orders of 73 asthmatic and 23 non-asthmatic patients’ capnogram were calculated. Based on the findings in this research, capnogram signals can be categorised as wide-sense nonstationary random signals. So, non-stationary techniques including linear predictive analysis and Burg algorithm analysis are used to process the capnogram signals. It should be noted that these techniques by windowing signal are based on this assumption that signal is locally stationary. Then, by means of Receiver Operating Characteristic (ROC) curve, the effectiveness of the extracted features to differentiate between asthmatic and non-asthmatic conditions is justified. Finally, selected features are used in a Gaussian radial basis function (GRBF) neural network. The output of this network is an integer prognostic index from 1 to 10 (depending on the severity of asthma) with an average good detection rate of 90.15% and an error rate of 9.85%. This developed algorithm is aspired to provide a fast and low-cost diagnostic system to help healthcare professional involved in respiratory care as it would be possible to monitor the severity of asthma automatically and instantaneously