Development of automatic retinal vessel segmentation method in fundus images via convolutional neural networks.

The analysis of fundus photograph is one of useful diagnosis tools for diverse retinal diseases such as diabetic retinopathy and hypertensive retinopathy. Specifically, the morphology of retinal vessels in patients is used as a measure of classification in retinal diseases and the automatic processing of fundus image has been investigated widely for diagnostic efficiency. The automatic segmentation of retinal vessels is essential and needs to precede computer-aided diagnosis system. In this study, we propose the method which implements patch-based pixel-wise segmentation with convolutional neural networks (CNNs) in fundus images for automatic retinal vessel segmentation. We construct the network composed of several modules which include convolutional layers and upsampling layers. Feature maps are made by modules and concatenated into a single feature map to capture coarse and fine structures of vessel simultaneously. The concatenated feature map is followed by a convolutional layer for performing a pixel-wise prediction. The performance of the proposed method is measured on DRIVE dataset. We show that our method is comparable to the results of other state-of-the-art algorithms.

Automated discrimination of dementia spectrum disorders using extreme learning machine and structural T1 MRI features.

The classification of neuroimaging data for the diagnosis of Alzheimer's Disease (AD) is one of the main research goals of the neuroscience and clinical fields. In this study, we performed extreme learning machine (ELM) classifier to discriminate the AD, mild cognitive impairment (MCI) from normal control (NC). We compared the performance of ELM with that of a linear kernel support vector machine (SVM) for 718 structural MRI images from Alzheimer's Disease Neuroimaging Initiative (ADNI) database. The data consisted of normal control, MCI converter (MCI-C), MCI non-converter (MCI-NC), and AD. We employed SVM-based recursive feature elimination (RFE-SVM) algorithm to find the optimal subset of features. In this study, we found that the RFE-SVM feature selection approach in combination with ELM shows the superior classification accuracy to that of linear kernel SVM for structural T1 MRI data.

Motor imagery EEG classification with optimal subset of wavelet based common spatial pattern and kernel extreme learning machine.

Performance of motor imagery based brain-computer interfaces (MI BCIs) greatly depends on how to extract the features. Various versions of filter-bank based common spatial pattern have been proposed and used in MI BCIs. Filter-bank based common spatial pattern has more number of features compared with original common spatial pattern. As the number of features increases, the MI BCIs using filter-bank based common spatial pattern can face overfitting problems. In this study, we used eigenvector centrality feature selection method, wavelet packet decomposition common spatial pattern, and kernel extreme learning machine to improve the performance of MI BCIs and avoid overfitting problems. Furthermore, the computational speed was improved by using kernel extreme learning machine.

Optimized automatic sleep stage classification using the normalized mutual information feature selection (NMIFS) method.

Sleep is a very important physiological phenomenon for recovery of physical and mental fatigue. Recently, there has been a lot of interest in the quality of sleep and the research is actively under way. In particular, it is important to have a repetitive and regular sleep cycle for good sleep. However, it takes a lot of time to determine sleep stages using physiological signals by experts. In this study, we constructed an optimized classifier based on normalized mutual information feature selection (NMIFS) and kernel based extreme learning machine (K-ELM), and total 4 sleep stages (Awake, weak sleep (stage1+stage2), deep sleep(stage3+stage4) and rapid eye movement (REM)) were automatically classified. As a results, the average of the accuracy obtained by proposed method (NMIFS+K-ELM) is 2~3% higher than that of simple method (K-ELM).

Changes in heart rate variability of patients with delirium in intensive care unit.

Delirium is an important syndrome in intensive care unit (ICU) patients, however, its characteristics are still unclear. Many evidences showed that this syndrome can be related to the autonomic instability. In this study, we aimed to investigate the possible alterations of autonomic nervous system (ANS) in delirium patients in ICU. Electrocardiography (ECG) of every ICU patient was measured during routine daily ICU care, and the data were gathered to evaluate the heart rate variability (HRV). HRV of total 60 patients were analyzed in time, frequency and non-linear domains. As a result, we found that heart rates of delirium patients were more variable and irregular than non-delirium patients. These findings may facilitate early detection and prevention of delirium in ICU.

Discrimination of multiple stress levels in virtual reality environments using heart rate variability.

People are suffering from various stress during daily living. Stress can cause a variety of symptoms, and in severe cases, it can lead to a dangerous disease. For this reason, it is essential to develop a simple method to evaluate stress level precisely. Popularly, heart rate variability (HRV) is used because it can reflect autonomic nervous system (ANS) activity. On the other hand, virtual reality (VR), which can provide environments similar to reality, is widely used in laboratory-based experiments. In this paper, we analyzed the HRV of healthy people by using the photoplethysmogram (PPG) while providing diverse stress situations. To detect and classify the exact stress levels, extracted HRV features and linear discriminant analysis (LDA) were utilized. As a result, high multi-class classification accuracy was obtained: Baseline (74%), mild stress (81%), and severe stress (82%).

A Preliminary Study on the Circadian Rhythm of Autonomic Nervous System in Patients Admitted to Intensive Care Unit, Using the Heart Rate Variability

OBJECTIVES: normal circadian rhythm of autonomic nervous system function stands for the daily change of sympathetic and parasympathetic modulation, which can be measured by heart rate variability (HRV). Generally, patients in the intensive care unit (ICU) are prone to sleep-wake cycle dysregulation, therefore, it may have an influence on the circadian rhythm of autonomic nervous system. This study was designed to interpret possible dysregulation of autonomic nervous system in ICU patients by using HRV. METHODS: HRV was assessed every 3 hours in 21 ICU patients during a 7-minute period. The statistical differences of HRV features between the morning (AM 6 : 00–PM 12 : 00), and the afternoon (PM 12 : 00–PM 18 : 00) periods were evaluated in time domain and frequency domain. RESULTS: Patients showed significantly increased normalized power of low frequencey (nLF), absolute power of low frequencey (LF)/absolute power of high frequencey (HF) in the afternoon period as compared to the morning period. However, normalized power of high frequency (nHF) was significantly decreased in the afternoon period. There was no statistically significant difference between the morning period and the afternoon period in the time domain analysis. CONCLUSIONS: The increased sympathetic tone in the afternoon period supports possible dysregulation in the circadian rhythm of autonomic nervous system in ICU patients. Future studies can help to interpret the association between autonomic dysregulation and negative outcomes of ICU patients.

Non-contact robust heart rate estimation using HSV color model and matrix-based IIR filter in the face video imaging.

The heart rate (HR) is one of the important indicators for observing the patient condition. Therefore, many estimation techniques for acquiring heart rate have been developed. The photoplethysmography (PPG) and electrocardiography (ECG) are the common measurement techniques for estimating the heart rate. However, they should contact on the human skin in order to estimate the accurate heart rate. In this paper, we propose a non-contact robust heart rate measurement method using a webcam device. In addition, we evaluated the performance of the proposed algorithm in comparison with other algorithms.

Tracking driver's heart rate by continuous-wave Doppler radar.

Developing driving safety system with medical assistance devices for preventing accidents has become a major social issue in recent year. These devices have been developed using electrocardiogram (ECG) and photoplethysmogram (PPG) for measuring the heart rate (HR). However, driver should directly contact with the sensor for monitoring the HR. Recently, non-contact system based on continuous-wave Doppler radar has widely studied for monitoring HR. The periodogram by Fast Fourier Transform (FFT) was used for estimating HR. However, if motion artifacts by movement of driver and vehicle vibration contaminate the radar signal, we cannot find spectral peak of HR using FFT. In this paper, we propose a method using multiple signal classification (MUSIC) for estimating HR. We compared MUSIC algorithms with a commonly used FFT method using real experiment data while driving. The results indicate that our proposed method can estimate HR accurately from received radar Doppler signal with motion artifacts.