A Brief on Snoring Data and Classification Methods

Abstract

Recently, research on snoring sound had gained interest especially in the area of classification in Obstructive Sleep Apnea (OSA) and distinction from non-snoring sounds. Classifiers such as Support Vector Machine (SVM), Recurrent Neural Network (RNN) and Convolutional Neural Network (CNN) had been used to meet this interest. These approaches relies on several underlying techniques such as Mel Frequency Cepstral Coefficients (MFCC), Short-Time Fourier transform (STFT) and several others to extract features from snore/apnea Spectrogram images before classification process. However, achieving desirable classification accuracy depends on choice of classifier, feature extraction techniques and, available dataset. In the paper, we presented a brief survey on existing methods and snore data acquisition processes to quickly expose and ease new researches in this domain to make appropriate choice from available methods.

Machine Learning and Answer Set Program Rules towards Traffic Light Management

“…In the future, plan to develop a smart approach and green IoT applications to enhance livability, workability, safety and sustainability, through assigning sensors that collect information about all roads constrains and uses smart devices or other systems to exchange data between traffic signals and vehicles. Also we will apply the modern deep learning classifier such as [28] and [29] to save time and more accuracy, In addition, solar energy can be used to feed controller and sensors, so this will reduce grid electricity consumption and pollution.…”