Kyulin Lee

Detection of Arousals in Patients with Respiratory Sleep Disorders Using a Single Channel EEG

Frequent arousals during sleep degrade the quality of sleep and result in sleep fragmentation. Visual inspection of physiological signals to detect the arousal events is inconvenient and time-consuming work. The purpose of this study was to develop an automatic algorithm to detect the arousal events. We proposed the automatic method to detect arousals based on time-frequency analysis and the support vector machine (SVM) classifier using a single channel sleep electroencephalogram (EEG). The performance of our method has been assessed using polysomnographic (PSG) recordings of nine patients with sleep apnea, snoring and excessive daytime sleepiness (EDS). By the proposed method, we could obtain sensitivity of 87.92% and specificity of 95.56% for the training sets, and sensitivity of 75.26% and specificity of 93.08% for the testing sets, respectively. We have shown that proposed method was effective for detecting the arousal events

The optimum coating condition by response surface methodology for maximizing vapor-permeable water resistance and minimizing frictional sound of combat uniform fabric

The objectives of this study are to investigate how the variables of the water-repellent coating condition, concentration of polyurethane (PU) and curing temperature, set up by response surface methodology, affect vapor-permeable water resistance and fabric frictional sound. Also it aims to analyze the relationship between tensile properties and the sound pressure level (SPL) of the fabric and, finally, to suggest the optimum coating condition for minimizing the fabric frictional sound and maximizing the vapor-permeable water resistance. It was observed that the higher PU concentration increased the water resistance and SPL, but decreased WVT (water vapor transmission). It was shown that higher curing temperature, the other variable of the coating condition, increased the water resistance and SPL but decreased WVT. The relationship between tensile properties and SPL was analyzed and it was found that tensile stress at break (R2 = .716) and toughness (R2 = .717) were highly related to SPL; however, tensile strain at break (R2 = .508) was not. Finally, the optimum coating condition for minimizing fabric frictional sound and maximizing vapor-permeable water resistance was obtained at the PU concentration of 60% and the curing temperature of 149.7℃, and the predicted SPL and WVT were 72.27 dB and 8478.85 g/m2 24 h, respectively. The coefficients of determination (R2) were 0.82 and 0.85, respectively, which indicate that the model fit was highly significant (p < 0.05).

Arrhythmia Classification with Reduced Features by Linear Discriminant Analysis

In this study, we proposed 17 input features based on wavelet coefficients for arrhythmia detection and, by applying linear discriminant analysis to these, reduced the feature dimension to be 4. Then, with newly constructed 4 dimension input feature, a multi-layer perceptrons classifier was tried to detect 6 types of arrhythmia beats. For evaluation of input features by linear discriminant analysis, the arrhythmia detection efficiency with these (LDA) was compared to that with original input features (ORG) and that with of input features by principle component analysis (PCA) respectively. When LDA was compared to ORG, the former showed similar or a little higher values than the latter for different types of arrhythmia beats except SVT. And, LDA showed to be persistently higher than PCA. By theses cross-validations, for the detection of several types of arrhythmia beats, the reduction of input feature dimension by linear discriminant analysis was revealed to be prior to that by principle component analysis. Even if LDA was compared to ORG, it maintained the acceptable level efficiency so that the time and computational costs would be expected to be cutdown dramatically. Finally, by the proposed algorithm, we could obtain the good accuracy of arrhythmia detection and that of NSR, SVT, PVC and VF was 99.52%, 99.43%, 98.59% and 99.88%, respectively

Classification of Heartbeats based on Linear Discriminant Analysis and Artificial Neural Network

In this paper, we proposed a heartbeat classification algorithm based on linear discriminant analysis and artificial neural network. For the input of classifier, we extracted 275 input features from the first derivative signal of ECG signal and RR interval information and it was reduced to be 6 by LDA. To evaluate the performance of the proposed algorithm, we compared the result of the proposed algorithm with that of fuzzy inference system classifier. MIT-BIH Arrhythmia database were used as test and learning data. The performance of the proposed algorithm was 97.49% for sensitivity, 97.91% for specificity and 96.36% for accuracy. For the extraction of features, the first derivative signal of ECG is used only so that the real-time implementation of this algorithm was possible. And, on account of the reduction of feature dimensionality, the time cost for learning and testing can be expected

Basic and Mechanical Properties by Film Type to Minimize the Sound Pressure Level of PTFE Laminated Vapor-permeable Water-repellent Fabrics

This study investigates the sound properties of fabric frictional sound (SPL, ∆L, ∆f) according to the film type of PTFE laminated vapor-permeable water-repellent fabrics in order to understand the relationship between SPL and the basic properties of fabrics such as layer, yarn type, and thickness of fiber. This study accesses their mechanical properties and determines how to control them to minimize SPL. Eight PTFE laminated water-repellent fabrics, composed of four different film types (A, B, C, D) and with two different fabrics, were used as test specimens. Frictional sounds generated at 1.21m/s were recorded by using a fabric sound generator and SPLs were analyzed through Fast Fourier Transformation (FFT). The mechanical properties of fabrics were measured by KES-FB. The SPL value was lowest at 74.4dB in film type A and highest as 85.5dB in type D. Based on ANOVA and post-hoc test, specimens were classified into less Loud Group (A, B) and Loud Group (C, D). It was shown that SPL was lower when 2 layer (instead of 3 layer), filament yarn than staple, and thin fiber than thick were used. In Group I, shearing properties (G, 2HG5), geometrical roughness (SMD), compressional properties (LC, RC) and weight (W) showed high correlation with SPL however, elongation (EM) and shear stiffness (G) did with SPL in Group II.

Detection of EEG Arousals in Patients with Respiratory Sleep Disorder

This paper describes the detection of arousals from sleep in patients with respiratory sleep disorder using a single channel EEG signal and support vector machine classifier. Determining the occurrence and the frequency of occurrence of arousals from sleep is very important because it is directly related to the quality of sleep. In this paper we used twenty polysomnographic recordings of patients with respiratory sleep disorder. Six recordings were used as training sets and fourteen recordings were used as test sets. We extracted three types of features, which are six indices relating to sleep states, the powers of each of four frequency bands and variations of power of EEG frequency, using time-frequency analysis. We detected arousals from sleep using the above features and SVM classifier. From the results, the sensitivity of 79.65% and the specificity of 89.52% were obtained. The error between the total arousal time detected by the proposed method and the annotated data was 15.09±10.76 min and it showed the possibility of application for the detection of arousal from sleep using a single channel EEG signal.

Development of a Fabric Friction Sound Generator Simulating Body Movement and Evaluation of the Generated Sound

Abstract: To investigate the sound generated by fabric friction, which simulates real wear conditions, a ‘fabric frictionsound generator’, which simulates body movement was developed. Fabric sounds from three specimens were gener-ated by the fabric sound simulator and recorded using high performance microphones. Physical sound parameters suchas sound pressure level (SPL), level range (∆L), and frequency difference (∆f) were calculated for the fabrics. All thephysical parameters (SPL, ∆L, and ∆f) of fabric sounds generated by the fourth-generation apparatus had lower valuescompared to the values obtained with the third-generation apparatus. Unlike the third-generation system, which gener-ates fabric sounds by reciprocating friction, the fourth-generation system was designed with silicon-based arm-and-legshaped abraders so that the levels of noise and fabric sounds generated were lower at all speeds. Keywords: fabric sound, simulating body movement, physical sound properties, sound pressure level (SPL), level range (∆L),frequency difference (∆f)

Estimation and Removal of the Stepped Deflation Artefact (SDA) in Oscillometric Blood Pressure Measurement

A new method to detect the blood pressure pulsation signal in oscillometric blood pressure measurement using stepped deflation method was developed. In stepped deflation method, morphology of the artefact signal changes much by bleed valve switching scenario. Various types of interference caused by the bleed valve switching are analyzed. The Adaptive Impulse Correlated Filter (AICF) was applied to remove the stepped deflation artefact (SDA). This AICF is performed using a parametric method which uses Finite Impulse Response (FIR) structure. The method is implemented using a specially designed acquisition system for this research. The analysis showed that amplitude and power of artefact increased with pulse-width of the bleed valve control pulse. The removal of artefact was almost accomplished without reducing the information in the blood pressure pulsation signal.