Snore Detection Research Articles

SnoreNet: Detecting Snore Events from Raw Sound Recordings.

Snoring is one of the earliest symptoms of Obstructive Sleep Apnea Hypopnea Syndrome (OSAHS). Snore detection is the first step in developing non-invasive, low-cost, and totally sound-based OSAHS analysis approaches. In this work, we propose a simple yet effective deep neural network, named SnoreNet, for detecting snores from a continuous sound recording. Without manually crafted features, SnoreNet can capture the characteristics of snores. Since snore varies in temporal length, SnoreNet combines output from multiple feature maps to detect snore. In each feature map, SnoreNet uses a set of default bounding box generated by a base length and different scales to match snores. SnoreNet adjusts the box to better locate snores and predicts a score for the presence of snore in each default bounding box. The performance of SnoreNet was evaluated on a newly collected snore pattern classes dataset, which achieves 81.82% average precision (AP).

0449 Comparing Deep Feature Representations to Improve Robustness to Subject Variation in Snore Detection

0449 Comparing Deep Feature Representations to Improve Robustness to Subject Variation in Snore Detection

Accelerometer based Active Snore Detection for Behavioral Modification.

Habitual snoring has been known to increase the risk for serious health problems in addition to affecting the quality of others' sleep. Several recent consumer products aim to automatically detect snoring events and wake the snorer to elicit a posture change. In this paper, we present a study comparing two of the methods, electromyography vs. accelerometry, proposed for automated snoring detection and incorporation of these into a wearable system. The study includes (a) the testing of various sensor configurations and placements to obtain optimal electromyography and accelerometry signals, (b) a review of the accuracy of a variety of snore detection algorithms from previously attained biological signals, and (3) design of an embedded device with integrated sensors and haptic feedback capability. Our preliminary results indicate superiority of accelerometry over electromyography. Further research opportunities to prove the concept and improve the design are then detailed for future work.

Recurrent Neural Network for Classification of Snoring and Non-Snoring Sound Events.

Obstructive sleep apnea (OSA) is a disorder that affects up to 38% of the western population. It is characterized by repetitive episodes of partial or complete collapse of the upper airway during sleep. These episodes are almost always accompanied by loud snoring. Questionnaires such as STOP-BANG exploit snoring to screen for OSA. However, they are not quantitative and thus do not exploit its full potential. A method for automatic detection of snoring in whole-night recordings is required to enable its quantitative evaluation. In this study, we propose such a method. The centerpiece of the proposed method is a recurrent neural network for modeling of sequential data with variable length. Mel-frequency cepstral coefficients, which were extracted from snoring and non-snoring sound events, were used as inputs to the proposed network. A total of 20 subjects referred to clinical sleep recording were also recorded by a microphone that was placed 70 cm from the top end of the bed. These recordings were used to assess the performance of the proposed method. When it comes to the detection of snoring events, our results show that the proposed method has an accuracy of 95%, sensitivity of 92%, and specificity of 98%. In conclusion, our results suggest that the proposed method may improve the process of snoring detection and with that the process of OSA screening. Follow-up clinical studies are required to confirm this potential.

Acoustics of snoring and automatic snore sound detection in children

Objective: Acoustic analyses of snoring sounds have been used to objectively assess snoring and applied in various clinical problems for adult patients. Such studies require highly automatized tools to analyze the sound recordings of the whole night’s sleep, in order to extract clinically relevant snore- related statistics. The existing techniques and software used for adults are not efficiently applicable to snoring sounds in children, basically because of different acoustic signal properties. In this paper, we present a broad range of acoustic characteristics of snoring sounds in children (N = 38) in comparison to adult (N = 30) patients. Approach: Acoustic characteristics of the signals were calculated, including frequency domain representations, spectrogram-based characteristics, spectral envelope analysis, formant structures and loudness of the snoring sounds. Main results: We observed significant differences in spectral features, formant structures and loudness of the snoring signals of children compared to adults that may arise from the diversity of the upper airway anatomy as the principal determinant of the snore sound generation mechanism. Furthermore, based on the specific audio features of snoring children, we proposed a novel algorithm for the automatic detection of snoring sounds from ambient acoustic data specifically in a pediatric population. The respiratory sounds were recorded using a pair of microphones and a multi-channel data acquisition system simultaneously with full-night polysomnography during sleep. Brief sound chunks of 0.5 s were classified as either belonging to a snoring event or not with a multi-layer perceptron, which was trained in a supervised fashion using stochastic gradient descent on a large hand-labeled dataset using frequency domain features. Significance: The method proposed here has been used to extract snore-related statistics that can be calculated from the detected snore episodes for the whole night’s sleep, including number of snore episodes (total snoring time), ratio of snore to whole sleep time, variation of snoring rate, regularity of snoring episodes in time and amplitude and snore loudness. These statistics will ultimately serve as a clinical tool providing information for the objective evaluation of snoring for several clinical applications.

Detection of Snores Using Source Separation on an Emfit Signal.

Snoring (SN) is an early sign of upper airway dysfunction, and it is strongly associated with obstructive sleep apnea. SN detection is important to monitor SN objectively and to improve the diagnostic sensitivity of sleep-disordered breathing. In this study, an automatic snore detection method using an electromechanical film transducer (Emfit) signal is presented. Representative polysomnographs of normal breathing and SN periods from 30 subjects were selected. Individual SN events were identified using source separation applying nonnegative matrix factorization deconvolution. The algorithm was evaluated using manual annotation of the polysomnographic recordings. According to our results, the sensitivity, and the positive predictive value of the developed method to reveal snoring from the Emfit signal were 82.81% and 86.29%, respectively. Compared to other approaches, our method adapts to the individual spectral snoring profile of the subject rather than matching a particular spectral profile, estimates the snoring intensity, and obtains the specific spectral profile of the snores in the epoch. Additionally, no training is necessary. This study suggests that it is possible to detect individual SN events with Emfit mattress, which can be used as a contactless alternative to more conventional methods such as piezo-snore sensors or microphones.

Validation of a new snoring detection device based on a hysteresis extraction algorithm

ObjectiveThis paper aims to introduce and validate our newly developed snoring detection device to automatically identify the incidence and amplitude of snores using the hysteresis extraction method. MethodsThirty patients (16 males and 14 females) with a history of snoring were included in this study. Each patient underwent a conventional polysomnography (PSG). Natural overnight snoring was recorded from each subject using our original snore detection device and an integrated circuit (IC) recorder while the patient slept during PSG. A new algorithm based on hysteresis extraction was used to detect snores and qualify the level of each event at 30-s intervals (one epoch). The automated and subjective assessment concordance was evaluated by comparing a total of 27,295 epochs, and sensitivity, specificity, and accuracy were calculated. ResultsStudy population analysis revealed a mean rate of snore time against the total sleep time of 14.1±7.9%. Further, validation of the automatic snore detection revealed the following: sensitivity, 71.2%; specificity, 93.1%; positive predictive value, 77.7%; negative predictive value, 94.6%; and accuracy, 90.7%. ConclusionsThis study revealed the efficacy of our newly developed snoring detection device and indicated that it may serve as a useful method in further snoring analysis via objective medical assessment. However, the sample size of 30 subjects was relatively small; therefore, further research is needed to evaluate this device.

Automatic snoring sounds detection from sleep sounds via multi-features analysis.

Obstructive sleep apnea hypopnea syndrome (OSAHS) is a serious respiratory disorder. Snoring is the most intuitively characteristic symptom of OSAHS. Recently, many studies have attempted to develop snore analysis technology for diagnosing OSAHS. The preliminary and essential step in such diagnosis is to automatically segment snoring sounds from original sleep sounds. This study presents an automatic snoring detection algorithm that detects potential snoring episodes using an adaptive effective-value threshold method, linear and nonlinear feature extraction using maximum power ratio, sum of positive/negative amplitudes, 500Hz power ratio, spectral entropy (SE) and sample entropy (SampEn), and automatic snore/nonsnore classification using a support vector machine. The results show that SampEn provides higher classification accuracy than SE. Furthermore, the proposed automatic detection method achieved over 94.0% accuracy when identifying snoring and nonsnoring sounds despite using small training sets. The sensitivity and accuracy of the results demonstrate that the proposed snoring detection method can effectively classify snoring and nonsnoring sounds, thus enabling the automatic detection of snoring.

Snoring sound classification from respiratory signal.

Snoring is common in the general population and the irregularity could lead to the presence of Obstructive sleep apnea. Diagnosis of OSA could therefore be made by snoring sound analysis. However, there is still a shortage of robust methods to automatically detect snoring sounds without the need to calibrate for every individual. In this paper, a novel method based on neural network is proposed to classify breathing sound episodes from snoring and non-snoring sound segments. Our snore detection algorithm was applied to the tracheal sounds of nine individuals with different OSA severities. On the testing dataset, the classifier achieved a sensitivity and specificity of 95.9% and 97.6% respectively. Our results indicate that using such a method could help to detect snoring sounds with high accuracy which would be useful in the diagnosis of sleep apnea.

압전센서를 이용한 코골이와 심박 검출을 위한 자동 알고리즘

In this paper, we proposed a novel method for automatic detection for snoring and heart beat using a single piezoelectric sensor. For this study multi-rate signal processing technique was applied to detect snoring and heart beat from the single source signal. The sound event duration and intensity features were used to snore detection and heart beat was found by autocorrelation. The performance of the proposed method was evaluated on clinical database, which is the nocturnal piezoelectric snoring data of 30 patients that suffered obstructive sleep apnea. The method achieved sensitivity of 88.6%, specificity of 96.1% with accuracy of 95.6% for snoring and sensitivity of 94.1% and positive predictive value of 87.6% for heart beat, respectively. These results suggest that the proposed method can be a useful tool in sleep monitoring and sleep disordered breathing diagnosis.

Nasal pressure recordings for automatic snoring detection

This study presents a rule-based method for automated, real-time snoring detection using nasal pressure recordings during overnight sleep. Although nasal pressure recordings provide information regarding nocturnal breathing abnormalities in a polysomnography (PSG) study or continuous positive airway pressure (CPAP) system, an objective assessment of snoring detection using these nasal pressure recordings has not yet been reported in the literature. Nasal pressure recordings were obtained from 55 patients with obstructive sleep apnea. The PSG data were also recorded simultaneously to evaluate the proposed method. This rule-based method for automatic, real-time snoring detection employed preprocessing, short-time energy and the central difference method. Using this methodology, a sensitivity of 85.4% and a positive predictive value of 92.0% were achieved in all patients. Therefore, we concluded that the proposed method is a simple, portable and cost-effective tool for real-time snoring detection in PSG and CPAP systems that does not require acoustic analysis using a microphone.

Efficient snoring and breathing detection based on sub-band spectral statistics.

Snoring, a common symptom in the general population may indicate the presence of obstructive sleep apnea (OSA). In order to detect snoring events in sleep sound recordings, a novel method was proposed in this paper. The proposed method operates by analyzing the acoustic characteristics of the snoring sounds. Based on these acoustic properties, the feature vectors are obtained using the mean and standard deviation of the sub-band spectral energy. A support vector machine is then applied to perform the frame-based classification procedure. This method was demonstrated experimentally to be effective for snoring detection. The database for detection included full-night audio recordings from four individuals who acknowledged having snoring habits. The performance of the proposed method was evaluated by classifying different events (snoring, breathing and silence) from the sleep sound recordings and comparing the classification against ground truth. The proposed algorithm was able to achieve an accuracy of 99.61% for detecting snoring events, 99.16% for breathing, and 99.55% for silence.

Polyvinylidene fluoride sensor-based method for unconstrained snoring detection

We established and tested a snoring detection method using a polyvinylidene fluoride (PVDF) sensor for accurate, fast, and motion-artifact-robust monitoring of snoring events during sleep. Twenty patients with obstructive sleep apnea participated in this study. The PVDF sensor was located between a mattress cover and mattress, and the patients’ snoring signals were unconstrainedly measured with the sensor during polysomnography. The power ratio and peak frequency from the short-time Fourier transform were used to extract spectral features from the PVDF data. A support vector machine was applied to the spectral features to classify the data into either the snore or non-snore class. The performance of the method was assessed using manual labelling by three human observers as a reference. For event-by-event snoring detection, PVDF data that contained ‘snoring’ (SN), ‘snoring with movement’ (SM), and ‘normal breathing’ epochs were selected for each subject. As a result, the overall sensitivity and the positive predictive values were 94.6% and 97.5%, respectively, and there was no significant difference between the SN and SM results. The proposed method can be applied in both residential and ambulatory snoring monitoring systems.

Nocturnal snoring sound analysis in the diagnosis of obstructive sleep apnea in the Chinese Han population.

Loud snoring is one of the principle symptoms of obstructive sleep apnea (OSA). Snoring sound analysis is a potentially cost-effective, reliable alternative for the diagnosis of OSA. However, no investigation has determined the accuracy of snoring signal analysis for the diagnosis of OSA in the Chinese Han population. Therefore, we investigated whether whole-night snoring detection and analysis aids the diagnosis of OSA using a new snore analysis technique. Snoring sounds were recorded using a non-contact microphone and polysomnography (PSG) was performed simultaneously throughout the night. We randomly selected 30 subjects each from four groups based on the severity of OSA. The rhythm and frequency domain of the snoring signal were analyzed based on frequency energy endpoint detection (FEP) and the Earth mover's distance (EMD), for each subject to harvest the EMD-calculated Apnea-Hypopnea Index (AHIEMD). Finally, we compared the AHIEMD with the PSG-monitored AHI (AHIPSG). The accuracy of the AHIEMD compared with the AHIPSG was 96.7, 86.7, 86.7, and 96.7% in non-, mild, moderate, and severe OSA patients, respectively. AHIEMD was correlated with AHIPSG (r(2) = 0.950, p < 0.001). The area under the receiver operating characteristic curve values for OSA detection was 0.974, 0.957, and 0.997 for AHIEMD thresholds of 5, 15, and 30 events/h, respectively. Bland-Altman analysis revealed 91.7% agreement of AHIEMD with AHIPSG. This new method for identifying OSA by analyzing snoring is feasible and reliable in the Han population. The snoring sound-based technique appears to be a promising tool for OSA screening and diagnosis.

Preliminary study for the personal handheld device based snoring detection in ordinary sleep situation.

Snoring is one of the representative phenomena of the sleep disorder and detection of snoring is quite important for improving quality of daily human life. The purpose of this research is to define the noises of the ordinary sleep situation and to find its characteristics as a preliminary research of snoring detection. Differently from previous snoring researches, we use a built-in sound recording system of Smartphone for practical use in ordinary sleep condition, and recording was carried out in a general private bedroom. Especially, we designed the experimental protocol, including the various noises could be frequently occurred during sleep such as cough, music, talking, alarm, door open/close, fan, radio and footstep to make closer to the actual sleep circumstance. The sound data set was recorded during actual sleep from 10 normal subjects. Totally 44 snoring data set and 75-noise dataset is acquired and analyzed.

Audio signal analysis in combination with noncontact bio-motion data to successfully monitor snoring.

This paper proposes a novel algorithm for automatic detection of snoring in sleep by combining non-contact bio-motion data with audio data. The audio data is captured using low end Android Smartphones in a non-clinical environment to mimic a possible user-friendly commercial product for sleep audio monitoring. However snore detection becomes a more challenging problem as the recorded signal has lower quality compared to those recorded in clinical environment. To have an accurate classification of snore/non-snore, we first compare a range of commonly used features extracted from the audio signal to find the best subject-independent features. Thereafter, bio-motion data is used to further improve the classification accuracy by identifying episodes which contain high amounts of body movements. High body movement indicates that the subject is turning, coughing or leaving the bed; during these instances snoring does not occur. The proposed algorithm is evaluated using the data recorded over 25 sessions from 7 healthy subjects who are suspected to be regular snorers. Our experimental results showed that the best subject-independent features for snore/non-snore classification are the energy of frequency band 3150-3650 Hz, zero crossing rate and 1st predictor coefficient of linear predictive coding. The proposed features yielded an average classification accuracy of 84.35%. The introduction of bio-motion data significantly improved the results by an average of 5.87% (p<;0.01). This work is the first study that successfully used bio-motion data to improve the accuracy of snore/non-snore classification.

Unconstrained snoring detection using a smartphone during ordinary sleep

BackgroundSnoring can be a representative symptom of a sleep disorder, and thus snoring detection is quite important to improving the quality of an individual’s daily life. The purpose of this research is to develop an unconstrained snoring detection technique that can be integrated into a smartphone application. In contrast with previous studies, we developed a practical technique for snoring detection during ordinary sleep by using the built-in sound recording system of a smartphone, and the recording was carried out in a standard private bedroom.MethodThe experimental protocol was designed to include a variety of actions that frequently produce noise (including coughing, playing music, talking, rining an alarm, opening/closing doors, running a fan, playing the radio, and walking) in order to accurately recreate the actual circumstances during sleep. The sound data were recorded for 10 individuals during actual sleep. In total, 44 snoring data sets and 75 noise datasets were acquired. The algorithm uses formant analysis to examine sound features according to the frequency and magnitude. Then, a quadratic classifier is used to distinguish snoring from non-snoring noises. Ten-fold cross validation was used to evaluate the developed snoring detection methods, and validation was repeated 100 times randomly to improve statistical effectiveness.ResultsThe overall results showed that the proposed method is competitive with those from previous research. The proposed method presented 95.07% accuracy, 98.58% sensitivity, 94.62% specificity, and 70.38% positive predictivity.ConclusionThough there was a relatively high false positive rate, the results show the possibility for ubiquitous personal snoring detection through a smartphone application that takes into account data from normally occurring noises without training using preexisting data.

Automatic Detection of Whole Night Snoring Events Using Non-Contact Microphone

ObjectiveAlthough awareness of sleep disorders is increasing, limited information is available on whole night detection of snoring. Our study aimed to develop and validate a robust, high performance, and sensitive whole-night snore detector based on non-contact technology.DesignSounds during polysomnography (PSG) were recorded using a directional condenser microphone placed 1 m above the bed. An AdaBoost classifier was trained and validated on manually labeled snoring and non-snoring acoustic events.PatientsSixty-seven subjects (age 52.5±13.5 years, BMI 30.8±4.7 kg/m2, m/f 40/27) referred for PSG for obstructive sleep apnea diagnoses were prospectively and consecutively recruited. Twenty-five subjects were used for the design study; the validation study was blindly performed on the remaining forty-two subjects.Measurements and ResultsTo train the proposed sound detector, >76,600 acoustic episodes collected in the design study were manually classified by three scorers into snore and non-snore episodes (e.g., bedding noise, coughing, environmental). A feature selection process was applied to select the most discriminative features extracted from time and spectral domains. The average snore/non-snore detection rate (accuracy) for the design group was 98.4% based on a ten-fold cross-validation technique. When tested on the validation group, the average detection rate was 98.2% with sensitivity of 98.0% (snore as a snore) and specificity of 98.3% (noise as noise).ConclusionsAudio-based features extracted from time and spectral domains can accurately discriminate between snore and non-snore acoustic events. This audio analysis approach enables detection and analysis of snoring sounds from a full night in order to produce quantified measures for objective follow-up of patients.

How to measure snoring?

Recently, snoring, independent of sleep apnea has been reported to have serious health consequences including carotid atherosclerosis. Detection of snoring is currently dependent on limited, poorly defined methods both for registration and analysis. Our study aims to add knowledge on how to measure and analyze snoring for future studies relating snoring to important health outcomes. Only subjects reporting habitual snoring were included in the study. Subjects were assessed with full in-laboratory polysomnography (Embla A10, Natus Medical Inc). Snoring was assessed with two overhead microphones, one chest microphone (T3 device, NoxMedical), a piezoelectric vibration sensor and an accelerometer on the neck and vibration in the nasal cannula. Our preliminary findings of n = 8 snorers showed a high correlation between the measured noise level of the chest microphone and the average dB of the two overhead microphones with the majority of events within 3 dB of each other. The fundamental frequency of snore events was measured from 50–250 Hz by sound analysis. However the three vibration sensors (piezoelectric, accelerometer and cannula) could only measure a range from 0–100 Hz. Therefore they could not pick up all snore events. The cannula additionally had a high noise floor, allowing it to be maximally 67% sensitive to snore events. The piezoelectric sensor was more sensitive to postural effects than the accelerometer and showed a significant increase in measured power when the subject lay on the same side as the sensor was positioned. Sound measurement of snoring is the most accurate objective analysis of snoring. Both cannula and neck vibration assessments of snoring have issues, causing them to miss out on a portion of snore events. Research support from the ResMed Foundation and the Science Fund of Landspitali University Hospital.

Study pattern of snoring and associated risk factors among medical students

Snoring can impair lifestyle and cause late cardiopulmonary complications. Early detection of snoring and timely intervention during adolescence can avoid complications. A single center prospective cross-sectional study was conducted in 548 undergraduate 17-25 years old undergraduate medical students of CSM Medical University, Lucknow, who were interviewed on the basis of the modified Berlin Questionnaire along with their room-partners describing their snoring habits. Ninety-seven (17.7%) subjects were found to be snorers. Risk factors viz. day time sleepiness (European Stroke Scale (ESS) scores), smoking habits, alcohol intake, neck size and BMI were observed. The proportion of males, smokers, BMI, neck size and ESS scores were significantly (p < 0.001) higher in snorers than non snorers. Logistic regression found sex (OR = 5.73, 95% CI = 2.11-15.61), smoking (OR = 2.20, 95% CI = 0.97-5.62), BMI (OR = 3.16, 95% CI = 1.09-7.36) and neck size (OR = 2.03, 95% CI = 0.79-6.35) as significant (p < 0.01) independent risk factors for snoring. A clinically significant form of Sleep Disordered Breathing (Habitual snoring, ESS score ≥ 11 and BMI ≥ 25 kg/m2) was suspected in 4 (0.7%) students. The findings of this study may also be validated in the general population. The study concluded that male undergraduate medical students are at a high risk for developing snoring habits.