|
Detection of sleep apnea through heart rate signal using Convolutional Neural Network
|
|
| Author:
|
, DR.SANTOSH KUMAR VISHWAKARMA, DR.SAURABH SINGH VERMA, MR.RAJIT NAIR, DR.VANDANA ROY4, AYUSHI AGRAWAL
|
| Abstract:
|
The most common type of various sleep-related breathing disorder is Obstructive Sleep Apnea (OSA). This is characterized by repeated sleep stoppages of the respiratory flow caused by the upper airway collapse. The OSA research protocol in sleep laboratories is primarily undiagnosed due to the impropriate polysomnography (PSG). This paper presents an automatic approach to the diagnosis of sleep apnea based on the heart rate signal. The work will predict or classify the heart signal based on sleep apnea or non- sleep apnea. We have implemented the deep learning model using Convolutional Neural Network with cross-validation to detect apnea and compare with all the existing machine learning algorithms such as logistic regression, multi-layer perceptron, and light GBM. The deep learning model CNN has computed 87% accuracy, a precision of 0.85, and a recall of 0.79, which is much better than other states of the art algorithms. The pytorch and CUDA is used for the implementation purpose with hyper tuning and Extra tree classifier for the feature extraction.
|
| Keyword:
|
EEG, Sleep Apnea, CNN, ECG.
|
| EOI:
|
-
|
| DOI:
|
https://doi.org/10.31838/ijpr/2020.12.04.654
|
| Download:
|
Request For Article
|
|
|