The world is currently going through the COVID-19 pandemic which has caused hundreds of thousands of deaths in just a few months. Considering the need for lockdown measures, most countries, including Malaysia, have implemented 'Movement Control Orders' (MCOs) as a prevention step to reduce the deadly spread of this disease. Local and worldwide media have reported the immediate improvement of air quality due to this event. Nevertheless, data on the effects of MCOs on air quality at local scales are still sparse. Here, we investigate changes in air quality during the MCO at an urban area using the air sensor network AiRBOXSense which measures monoxide (CO) and particulate matter (PM 2.5 and PM 10). In this study, air pollutant data during normal days were compared with MCO days using a reference analyser and AiRBOXSense. The results showed that the levels of the measured pollutants dropped by ~20 to 60% during the MCO days at most locations. However, CO in Kota Damansara (KD) dropped to 48.7%, but PM 2.5 and PM 10 increased up to 60% and 9.7% respectively during MCO days. Local burning activities in the residential area of KD are believed to be the main cause of the increased PM levels. This study has proven that air pollutant levels have significantly fallen due to the MCO. This air quality level information showed that the reduction of air pollutants can be achieved if traffic and industry emissions are strictly controlled.

The early detection and classification of dementia are important clinical support tasks for medical practitioners in customizing patient treatment programs to better manage the development and progression of these diseases. Efforts are being made to diagnose these neurodegenerative disorders in the early stages. Indeed, early diagnosis helps patients to obtain the maximum treatment benefit before significant mental decline occurs. The use of electroencephalogram as a tool for the detection of changes in brain activities and clinical diagnosis is becoming increasingly popular for its capabilities in quantifying changes in brain degeneration in dementia. This paper reviews the role of electroencephalogram as a biomarker based on signal processing to detect dementia in early stages and classify its severity. The review starts with a discussion of dementia types and cognitive spectrum followed by the presentation of the effective preprocessing denoising to eliminate possible artifacts. It continues with a description of feature extraction by using linear and nonlinear techniques, and it ends with a brief explanation of vast variety of separation techniques to classify EEG signals. This paper also provides an idea from the most popular studies that may help in diagnosing dementia in early stages and classifying through electroencephalogram signal processing and analysis.

Plants are a major source of food for the world population. Plant diseases contribute to production loss, which can be tackled with continuous monitoring. Manual plant disease monitoring is both laborious and error-prone. Early detection of plant diseases using computer vision and artificial intelligence (AI) can help to reduce the adverse effects of diseases and also overcome the shortcomings of continuous human monitoring. In this work, we propose the use of a deep learning architecture based on a recent convolutional neural network called EfficientNet on 18,161 plain and segmented tomato leaf images to classify tomato diseases. The performance of two segmentation models i.e., U-net and Modified U-net, for the segmentation of leaves is reported. The comparative performance of the models for binary classification (healthy and unhealthy leaves), six-class classification (healthy and various groups of diseased leaves), and ten-class classification (healthy and various types of unhealthy leaves) are also reported. The modified U-net segmentation model showed accuracy, IoU, and Dice score of 98.66%, 98.5%, and 98.73%, respectively, for the segmentation of leaf images. EfficientNet-B7 showed superior performance for the binary classification and six-class classification using segmented images with an accuracy of 99.95% and 99.12%, respectively. Finally, EfficientNet-B4 achieved an accuracy of 99.89% for ten-class classification using segmented images. It can be concluded that all the architectures performed better in classifying the diseases when trained with deeper networks on segmented images. The performance of each of the experimental studies reported in this work outperforms the existing literature.

Diabetes mellitus (DM) is one of the most prevalent diseases in the world, and is correlated to a high index of mortality. One of its major complications is diabetic foot, leading to plantar ulcers, amputation, and death. Several studies report that a thermogram helps to detect changes in the plantar temperature of the foot, which may lead to a higher risk of ulceration. However, in diabetic patients, the distribution of plantar temperature does not follow a standard pattern, thereby making it difficult to quantify the changes. The abnormal temperature distribution in infrared (IR) foot thermogram images can be used for the early detection of diabetic foot before ulceration to avoid complications. There is no machine learning-based technique reported in the literature to classify these thermograms based on the severity of diabetic foot complications. This paper uses an available labeled diabetic thermogram dataset and uses the k-mean clustering technique to cluster the severity risk of diabetic foot ulcers using an unsupervised approach. Using the plantar foot temperature, the new clustered dataset is verified by expert medical doctors in terms of risk for the development of foot ulcers. The newly labeled dataset is then investigated in terms of robustness to be classified by any machine learning network. Classical machine learning algorithms with feature engineering and a convolutional neural network (CNN) with image-enhancement techniques are investigated to provide the best-performing network in classifying thermograms based on severity. It is found that the popular VGG 19 CNN model shows an accuracy, precision, sensitivity, F1-score, and specificity of 95.08%, 95.08%, 95.09%, 95.08%, and 97.2%, respectively, in the stratification of severity. A stacking classifier is proposed using extracted features of the thermogram, which is created using the trained gradient boost classifier, XGBoost classifier, and random forest classifier. This provides a comparable performance of 94.47%, 94.45%, 94.47%, 94.43%, and 93.25% for accuracy, precision, sensitivity, F1-score, and specificity, respectively.

Respiratory ailments such as asthma, chronic obstructive pulmonary disease (COPD), pneumonia, and lung cancer are life-threatening. Respiration rate (RR) is a vital indicator of the wellness of a patient. Continuous monitoring of RR can provide early indication and thereby save lives. However, a real-time continuous RR monitoring facility is only available at the intensive care unit (ICU) due to the size and cost of the equipment. Recent researches have proposed Photoplethysmogram (PPG) and/ Electrocardiogram (ECG) signals for RR estimation however, the usage of ECG is limited due to the unavailability of it in wearable devices. Due to the advent of wearable smartwatches with built-in PPG sensors, it is now being considered for continuous monitoring of RR. This paper describes a novel approach for RR estimation using motion artifact correction and machine learning (ML) models with the PPG signal features. Feature selection algorithms were used to reduce computational complexity and the chance of overfitting. The best ML model and the best feature selection algorithm combination was fine-tuned to optimize its performance using hyperparameter optimization. Gaussian Process Regression (GPR) with Fit a Gaussian process regression model (Fitrgp) feature selection algorithm outperformed all other combinations and exhibits a root mean squared error (RMSE), mean absolute error (MAE), and two-standard deviation (2SD) of 2.63, 1.97, and 5.25 breaths per minute, respectively. Patients would be able to track RR at a lower cost and with less inconvenience if RR can be extracted efficiently and reliably from the PPG signal.

Conventional air quality monitoring systems, such as gas analysers, are commonly used in many developed and developing countries to monitor air quality. However, these techniques have high costs associated with both installation and maintenance. One possible solution to complement these techniques is the application of low-cost air quality sensors (LAQSs), which have the potential to give higher spatial and temporal data of gas pollutants with high precision and accuracy. In this paper, we present DiracSense, a custom-made LAQS that monitors the gas pollutants ozone (O3), nitrogen dioxide (NO2), and carbon monoxide (CO). The aim of this study is to investigate its performance based on laboratory calibration and field experiments. Several model calibrations were developed to improve the accuracy and performance of the LAQS. Laboratory calibrations were carried out to determine the zero offset and sensitivities of each sensor. The results showed that the sensor performed with a highly linear correlation with the reference instrument with a response-time range from 0.5 to 1.7 min. The performance of several calibration models including a calibrated simple equation and supervised learning algorithms (adaptive neuro-fuzzy inference system or ANFIS and the multilayer feed-forward perceptron or MLP) were compared. The field calibration focused on O3 measurements due to the lack of a reference instrument for CO and NO2. Combinations of inputs were evaluated during the development of the supervised learning algorithm. The validation results demonstrated that the ANFIS model with four inputs (WE OX, AE OX, T, and NO2) had the lowest error in terms of statistical performance and the highest correlation coefficients with respect to the reference instrument (0.8 < r < 0.95). These results suggest that the ANFIS model is promising as a calibration tool since it has the capability to improve the accuracy and performance of the low-cost electrochemical sensor.

In this paper, a first-of-a-kind mixed-mode universal filter employing three VDBAs and three passive components, is proposed. The filter operates in all four modes and provides all five filter responses, namely voltage-mode (VM), current-mode (CM), trans-impedance-mode (TIM), or trans-admittance-mode (TAM). Additionally, the same filter topology can also work as a CM single-input-multi-output (SIMO) filter. A state-of-the-art comparison of various ‘voltage differencing’ variants of the voltage differencing buffered amplifier (VDBA)-based SIMO/MISO (single-input-multi-output/multi-input-single-output)-type biquad filters further highlight the significance of the presented research. In the proposed no passive component matching is required for generating the filter responses. The filter circuit also provides inbuilt tunability of the quality factor independent of the pole frequency. The non-ideal frequency dependent gain and component sensitivity analyses of the filter were also performed. The Silterra Malaysia 0.18μm process design kit (PDK) is employed to design and validated the proposed VDBA-based filter using the Cadence design software. The simulation results closely follow the theoretical predictions. To further verify the practical feasibility of the proposed filter, an experimental evaluation is also completed. The VDBA-based filter is implemented using off-the-shelf operational transconductance amplifiers Intersil CA3080, Texas Instruments LF356 op-amp, and Analog Devices AD844s. The filter is designed for a characteristic frequency of 100 kHz. The time and frequency domain measurement results indicate the proper functioning of the filter.

No abstract