Sleep snoring detection using multi-layer neural networks

Abstract

Abstract. Snoring detection is important for diagnosing obstructive sleep apnea syndrome (OSAS) and other respiratory sleep disorders. In general, audio signal processing such as snoring sound analysis uses the frequency characteristics of the signal. Recently, a correlational filter Multilayer Perceptron neural network (f-MLP) has been proposed, which has the first hidden layer of correlational filter operations in frequency domain. It demonstrated a superior classification performance for the pattern sets; of these, frequency information is the dominant feature for classification. The first hidden layer is implemented with the correlational filter operation; its output is the power spectrum of the filter output, while the other layers are the same as the ordinary multilayer Perceptron (o-MLP). By using the back-propagation learning algorithm for the correlational filter layer, f-MLP was able to self-adapt the filter coefficients to produce its output with more discrimination power for classification in the higher layer. In this research, this f-MLP was applied for sleep snoring signal detection. As a result, the f-MLP achieved an average detection rate of 96% for the test patterns, compared to the conventional multilayer neural network that demonstrates an 82% average detection rate.

Automatic snoring detection using a hybrid 1D–2D convolutional neural network

“…And these algorithms have yielded promising results in characterizing the nocturnal sleep respiratory audio signal subsequently. For instance, Nguyen et al 39 and Çavuşoğlu et al 40 respectively utilized multilayer perceptron neural networks (MLP) to differentiate between snore and non-snore sounds; Arsenali et al 41 applied a long-short term memory (LSTM) model to classify snoring and non-snoring sounds after extracting MFCCs; Sun et al 42 proposed SnoreNet, a one-dimensional CNN (1D CNN) that directly operates on raw sound signals without manually crafted features; Khan et al 8 developed a two-dimension CNN (2D CNN) to analyze MFCCs images for automatically detecting snoring and applied it into a wearable gadget; Jiang et al 43 found an optional combination of Mel-spectrogram and CNN-LSTM-DNN for snoring recognition; Xie et al 44 employed a CNN to extract features from the constant-Q transformation (CQT) spectrogram, and then a recurrent neural network (RNN) was utilized to process the sequential CNN output for classifying the audio signal to snore or non-snore events. The comprehensive experiment settings and results of these researches are provided in Table 1 .…”

Usefulness of Artificial Neural Networks in the Diagnosis and Treatment of Sleep Apnea-Hypopnea Syndrome

“…In a recent study, Nguyen and Won [54] proposed a novel correlational filter ANN (f-MLP) to distinguish normal breathing patterns from snoring patterns during sleep. This ANN implements a correlational filter operation in the frequency domain in a first hidden layer aimed at improving the discriminant power of the spectral content of input patterns, followed by a second feedforward hidden layer.…”

Recognition of Hand Gestures using Wavelet Packet Transform and Cascaded Feed Forward Neural Networks

“…FirasAlomariet.al [34] proposed a pattern recognition system using the energy of wavelet coefficients to categorise eight different hand gestures. In [35,36], an Artificial Neural Network (ANN) was trained using information on the level of muscular contraction and previous classifier outputs in order to determine the accuracy of the classifier's decision., YinaGuoet.al [37] used ANN to categorise six different hand movements and achieved excellent accuracy, however the system was not resilient when amputees' electrodes were shifted, changed in size, or oriented differently.…”