Design of Apnea Detection Interface Including Time Delay Neural Networks for Portable Recording Devices with Three Channels

Peker O., YILMAZ A., DEMİR A. U.

26th IEEE Signal Processing and Communications Applications Conference (SIU), İzmir, Turkey, 2 - 05 May 2018 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/siu.2018.8404838
  • City: İzmir
  • Country: Turkey
  • Hacettepe University Affiliated: Yes


In this study, a software platform was developed in order to examine physiological signals recorded by portable recording devices used for apnea prescreening purposes and collaborate them for identifying apnea episodes to decide whether standart but more complicated polysomnography test stage might be required for possible apnea patients. Thus, based on the third level portable apnea device definition, three channels, namely air flow, oxygen saturation and ECG signal, are used for apnea detections[13]. For the detection of apnea, neural network topologies can be choosen through our designed interface and associated features that are derived for each corresponding channels can be used for realizing training and testing phases. For a sample case, the neural network trained with hybrid noisy data generated from the records obtained by the portable recording device has shown 89% high detection rate based on the expert scores.