Outline

  • Abstract
  • I. Introduction
  • II. Proposed Method
  • III. Result and Discussion
  • IV. Conclusion

رئوس مطالب

  • چکیده
  • کلیدواژه ها
  • 1. مقدمه
  • 2. روش پیشنهاد شده
  • الف) استخراج سیگنال محدود به باند
  • ب)استخراج ویژگی از چند باند
  • ج) استخراج الگوی تغییرات ویژگی ها به صورت زمانی
  • د) برازش مدل از الگوهای تغییرات ویژگی ها
  • ه) دسته بندی کننده
  • 3. نتایج و مباحث
  • الف) دیتابیس ها
  • ب) نکویی برازش
  • ج) نکویی ویژگی ها
  • د) نتایج دسته بندی
  • 4. جمع بندی
  • منابع

Abstract

Sleep apnea, a serious sleep disorder affecting a large population, causes disruptions in breathing during sleep. In this paper, an automatic apnea detection scheme is proposed using single lead electroencephalography (EEC) signal to discriminate apnea patients and healthy subjects as well as to deal with the difficult task of classifying apnea and non apnea events of an apnea patient. A unique multiband subframe based feature extraction scheme is developed to capture the feature variation pattern within a frame of EEC data, which is shown to exhibit significantly different characteristics in apnea and nonapnea frames. Such within frame feature variation can be better represented by some statistical measures and characteristic probability density functions. It is found that use of Rician model parameters along with some statistical measures can offer very robust feature qualities in terms of standard performance criteria, such as Bhattacharyya distance and geometric separability index. For the purpose of classification, proposed features are used in K Nearest Neighbor classifier. From extensive experimentations and analysis on three different publicly available databases it is found that the proposed method offers superior classification performance in terms of sensitivity, specificity, and accuracy.

Keywords: - - - - - - - - -

Conclusions

In conventional frame-by-frame EEG data analysis only the global characteristics of a frame can be obtained as in that case, features are extracted considering the entire frame at a time. On the contrary, in this paper, two-stage feature extraction method is proposed. First, the feature is computed from small duration overlapping sub-frames within a frame, which can precisely capture sharp changes with respect to time and provide temporal variation of the extracted feature within that frame. Next, statistical analysis and modeling are carried out on the resulting feature variation pattern, which gives an opportunity to utilize both local and global characteristics of a frame. Apart from ensuring such time resolution in feature extraction, use of multi-band signals also ensures frequency resolution. Among various PDF models, it is found that the Rician PDF is offering the best feature quality in terms of Bhattacharyya distance and GSI. Irrespective of the type of apnea, the proposed method can not only classify apnea patient and healthy subject but also classify apnea and nonapnea frames of an apnea patient, which has a great demand in the overnight polysomnography (PSG) to reduce human error, labor and cost. The proposed method is evaluated on three different and large EEG databases and it offers superior classification performance in comparison to some existing methods in terms of sensitivity, specificity and accuracy. It makes the proposed method to be widely applicable in a greater domain of diagnosis.

دانلود ترجمه تخصصی این مقاله دانلود رایگان فایل pdf انگلیسی