International Journal of Computer Science & Emerging Technologies (E-ISSN: 2044-6004) 38 Volume 1, Issue 4, December 2010 Enhanced Method for Extracting Features of Respiratory Signals and Detection of Obstructive Sleep Apnea Using Threshold Based Automatic Classification Algorithm A.Bhavani Sankar 1 , D.Kumar 2 , K.Seethalakshmi 3 1 Assistant Professor, Dept. of ECE, Anjalai Ammal - Mahalingam Engineering College, Kovilvenni 2 Dean, Research,Periyar Maniyammai University,Vallam 3 Senior Lecturer, Dept. of ECE, Anjalai Ammal - Mahalingam Engineering College, Kovilvenni. E-Mail: absankar72@gmail.com, kumar_durai@yahoo.com, seetha.au@gmail.com Abstract — Obstructive Sleep Apnea is a frequent disorder with detrimental health, performance and safety effects. The diagnosis of the disorder is cumbersome and expensive. New methods for screening and diagnosis are needed. The method we describe in this work is based on detection of four main features of respiratory signal. The automatic signal classification starts by extracting signal features from a 1 minute data segment through autoregressive modeling (AR) and other techniques. Four features are: signal energy, zero crossing frequency, dominant frequency estimated by AR and strength of dominant frequency based on AR. These features are then compared to threshold values and introduced to a series of conditions to determine the signal category for each specific epoch. The threshold values for the parameters were determined through experiment. Keywords- Sleep Apnea, Motion Artifact, Energy Index, Respiration rate, Dominant frequency, Strength of Dominant frequency, Zero Crossing. 1. INTRODUCTION Respiration monitors are of crucial importance in providing timely information regarding pulmonary function in adults and the incidence of Sudden Infant Death Syndrome (SIDS) in neonates. However, to accurately monitor respiration, the noise inherent in measuring devices, as well as artifacts introduced by body movements must be removed or discounted. With the recent success of media in creating awareness about the importance of sleep and effects of sleep apnea the classifying algorithm should be easy to use and provide a fair prediction that must contribute to public health. One can imagine a multitude of intelligent classification algorithms that could help to reach better identification mechanism. For example an algorithm should be capable of classifying different types of signal with different characteristics feature. Such an algorithm has the potential to become major classification tool. There have been enormous growth in developing efficient algorithm for classification of the respiratory signals, the reduced computational steps, reduced number of parameters used, increasing the capability to differentiate the signals and easy to implement in hardware setup to provide clinical support. An efficient algorithm should adopt itself to any kind of signals; it should not have any static rules for classifying the given input signal. This work shows a simple method for respiratory signal classification using a MATLAB coding. It describes an automatic classification algorithm using features derived from the autoregressive modeling and threshold crossing schemes that was used to classify respiratory signals into the following categories: (1) normal respiration, (2) respiration with artifacts and (3) sleep apnea. This classification is capable of detecting fatigue of the human by identifying sleep apnea, early detection of sleep troubles and disorders in groups at risk, reduces the risks of being affected by serious heart diseases in future. The main contribution of this paper is the analysis of signals those are necessary for classification of the respiratory signals which yields not only the classification but also the analysis of various ailments. Results in [1] indicate that respiratory signals alone are sufficient and perform even better than the combined respiratory and ECG signals. Respiratory signals are convenient to measure because they do not require electrodes on the skin, and people may wear the sensors for periods of several days and weeks. An apnea detection method based on spectral analysis was discussed in detail in [2]. In [3] the possibility of recognizing obstructive sleep apnea based on beat-by-beat features in ECG recordings was studied. It was also explored the application of time- varying autoregressive models and KNN linear classifier. A classification scheme of respiratory signal based on fuzzy logic was proposed in [4]. The paper [5] proposes an implementation of automatic classification of respiratory signals using a Field Programmable Gate Array (FPGA). The main novelty in [6] is that the phase difference between the two respiration signals is considered in order to determine the presence and grade of obstructive apnea. The work in [7] shows that the interval