Search

The detection of microorganisms is important in numerous applications such as water quality monitoring, blood analysis, and food testing. The conventional detection methods are tedious and labour-intensive. Establish methods involve culturing, counting and identification of the pathogen by an experienced technician which typically can take several days. The use of opto-fluidic technology to capture microorganism images offers 0 route to reduce the overall assay time. However, the detection still requires a trained technician. This paper proposes an image processing method that can be used to classify microorganism images captured by an opto-fluidic set up in an automatic manner. The proposed algorithm incorporates some of the features used in other microorganism image detection methods and proposes two new features-Entropy of Histogram of Oriented Gradients (HOG) and the filtered intensities. In addition, we propose to apply the minimal-Redundancy-Maximal-Relevance (mRMR) criterion to select and rank these features. The probability and joint probability distribution functions of the mRMR are estimated using a Gaussian model and the Kernel Density Estimation model. The performance of the proposed method was validated using SVM and data collected from an experimental setup. The results show that our proposed method outperforms existing methods and is capable of achieving a classification accuracy up to 95.8%.

The current gold standard of Obstructive Sleep Apnea (OSA) diagnosis involves the use of a Polysomnography (PSG) system which requires the patient to stay in the hospital for overnight recording. The process is uncomfortable for the patient and it disturbs the patient's sleep pattern. On the other hand, it is well known that some acoustic features of the snoring sounds are good indicators of the presence of OSA, and a variety of acoustic OSA detection algorithms have been reported in the literature. Typically, these algorithms use multiple features and a relatively complex classifier, which are not ideal for handling the huge over-night data. In this paper, we propose an algorithm that uses a single feature and a relatively simple classifier. The proposed feature is the difference between two carefully selected Mel-frequency cepstral coefficients (MFCCs) of the snoring sound samples. The proposed classifier is derived based on a modified minimum distance criterion. The proposed algorithm has been tested with patient data. The results show that the proposed algorithm outperforms existing algorithms and is able to achieve up to 97.1% detection accuracy.

Obstructive Sleep Apnea (OSA) is characterized by repetitive episodes of airflow reduction (hypopnea) or cessation (apnea), which, as a prevalent sleep disorder, can cause people to stop breathing for 10 to 30 seconds at a time and lead to serious problems such as daytime fatigue, impaired memory, and depression. This work intends to explore automatic detection of OSA events with 1-second annotation based on blood oxygen saturation, oronasal airflow, and ribcage and abdomen movements. Deep Learning (DL) technology, specifically, Convolutional Neural Network (CNN), is employed as a feature detector to learn the characteristics of the highorder correlation among visible data and corresponding labels. A fully-connected layer in the last stage of the CNN is connected to the output layer and constructs the desired number of outputs for sleep apnea events classification. A leave-one-out cross-validation has been conducted on the PhysioNet Sleep Database provided by St. Vincents University Hospital and University College Dublin, and an average accuracy of $79 .61$% across normal, hypopnea, and apnea, classes is achieved.

Presence of wheezes in breathing sounds has been associated with several respiratory and pulmonary diseases. In this paper we present a novel low-complexity wheeze detection method based on frequency contour tracking for automatic wheeze detection. Two hardware friendly variants of the algorithm have also been proposed. Applying the proposed feature extraction algorithm we achieved very high classification accuracy (> 99%) at considerably low computational complexity (3×-6×) compared to earlier methods and the power consumption of the proposed method is shown to be significantly less (70×-100×) compared to `record and transmit' strategy in wearable devices.

Scent plays an important role in influencing the brain and has been commonly used in psychological research. Much of such research has been conducted without the use of electroencephalography (EEG) to measure the response of the human brain to scent stimulus. Recent studies have involved the use of EEG to perform comparative studies on how different scents can affect brain activity. However, little has been done to analyze the trend of brain activity when a subject is repeatedly exposed to the same scent. This paper discusses the use of 4 features - Entropy Difference, Entropy Ratio, Entropy Time and Root Mean Square (RMS) to perform trend analysis of EEG signals in a repeated scent-exposure setting. The results show that different types of scents cause the brain to be stimulated at different degrees for each repeated exposure, giving rise to different trend patterns. It is also observed that the 4 features give similar trends for the same scent. This similarity allows us to combine the 4 features by forming a feature vector and plotting them in 3 dimensional (3D) space, using 3 repeated scent exposures as the axes. The region of space where the feature vector lies is represented by an ellipsoid, which can be used to characterize a particular scent. Unlike previous work, which did not characterize scent from EEG recordings, this paper investigates the different trends of scent after its repeated exposure to the human subject and by using the 3D representation to characterize the scent.

The presence of an excessive amount of water in lung is a sign of pulmonary edema which can be caused by heart failure. The current solutions for lung water detection involve the use of X-ray, CT scan or serum biomarkers, which require bulky and expensive instruments as well as long measurement duration. This paper reports on a study conducted on the use of a different sensing modality to detect the presence of water in lung. The main contributions of the paper are twofold: 1) we propose to employ acoustic (or sound) based techniques for lung water detection. The design is simple and can be implemented on a portable or wearable system; 2) we establish the feasibility of sound-based techniques for lung water detection, by carrying out experimental studies using four feature extraction methods combined with two classification methods. The findings of this study will be beneficial to the design of portable devices for rapid and accurate lung water detection.

In this paper, we propose to use DWT coefficients as features for emotion recognition from EEG signals. Previous feature extraction methods used power spectra density values dervied from Fourier Transform or sub-band energy and entropy derived from Wavelet Transform. These feature extracion methods eliminate temporal information which are essential for analyzing EEG signals. The DWT coefficients represent the degree of correlation between the analyzed signal and the wavelet function at different instances of time; therefore, DWT coefficients contain temporal information of the analyzed signal. The proposed feature extraction method fully utilizes the simultaneous time-frequency analysis of DWT by preserving the temporal information in the DWT coefficients. In this paper, we also study the effects of using different wavelet functions (Coiflets, Daubechies and Symlets) on the performance of the emotion recognition system. The input EEG signals were obtained from two electrodes according to 10-20 system: F(p1) and F(p2). Visual stimuli from International Affective Picture System (IAPS) were used to induce two emotions: happy and sad. Two classifiers were used: Extreme Learning Machine (ELM) and Support Vector Machine (SVM). Experimental results confirmed that the proposed DWT coefficients method showed improvement of performance compared to previous methods.

The respiratory rate is a vital sign that can provide important information about the health of a patient, especially that of the respiratory system. The aim of this study is to develop a simple method that can be applied in wearable systems to monitor the respiratory rate automatically and continuously over extended periods of time. In this paper, a novel respiratory rate estimation method is presented to achieve this target. The proposed method has been evaluated in both the open-source data as well as the local-hospital data, and the results are encouraging. The findings of this study revealed strong linear correlation to the reference respiratory rate. The correlation coefficients for the open-source data and the in-hospital data are 0.99 and 0.96 respectively. The standard deviation of the estimation error is less than 7% for both types of data.