TRANSFER LEARNING METHOD FOR SLEEP STAGES CLASSIFICATION USING DIFFERENT DOMAIN
MOHAMMAD MAHDI MORADI
Department of Biomedical Engineering, Kazerun Branch, Islamic Azad University, Kazerun, Iran.
MOHAMMAD H. FATEHI *
Department of Electrical Engineering, Kazerun Branch, Islamic Azad University, Kazerun, Iran.
HASSAN MASOUMI
Department of Biomedical Engineering, Kazerun Branch, Islamic Azad University, Kazerun, Iran.
MEHDI TAGHIZADEH
Department of Electrical Engineering, Kazerun Branch, Islamic Azad University, Kazerun, Iran.
*Author to whom correspondence should be addressed.
Abstract
Sleep stages classification using the signal analysis includes electroencephalogram (EEG), Electrooculography (EOG), Electromyography (EMG), Photoplethysmogram (PPG), and electrocardiogram (ECG). In this study, the proposed method using transfer learning to sleep stages classification. First, we have used the PPG and ECG signals, because they are less complex. This signal has the least complexity, and in this article we used this signal for transitional learning. n this study, we extracted 52 features from two signals and prepared for the classification stage. This method includes two steps, (a) Train data PPG and Test data ECG, (b) Train data ECG and Test data PPG. Results proved that our method has acceptable reliability for classification. The accuracy of 95.25% and 94.63% has been reached.
Keywords: PPG signal, ECG signal, sleep stages classification, transfer learning.