The human body can act as a medium for the transmission of electromagnetic waves in the wireless body sensor networks context. However, there are transmission losses in biological tissues due to the presence of water and salts. This Letter focuses on lateral intra-body microwave communication through different biological tissue layers and demonstrates the effect of the tissue thicknesses by comparing signal coupling in the channel. For this work, the authors utilise the R-band frequencies since it overlaps the industrial, scientific and medical radio (ISM) band. The channel model in human tissues is proposed based on electromagnetic simulations, validated using equivalent phantom and ex-vivo measurements. The phantom and ex-vivo measurements are compared with simulation modelling. The results show that electromagnetic communication is feasible in the adipose tissue layer with a low attenuation of ∼2 dB per 20 mm for phantom measurements and 4 dB per 20 mm for ex-vivo measurements at 2 GHz. Since the dielectric losses of human adipose tissues are almost half of ex-vivo tissue, an attenuation of around 3 dB per 20 mm is expected. The results show that human adipose tissue can be used as an intra-body communication channel.

Low-power wearable devices for disease diagnosis are used at anytime and anywhere. These are non-invasive and pain-free for the better quality of life. However, these devices are resource constrained in terms of memory and processing capability. Memory constraint allows these devices to store a limited number of patterns and processing constraint provides delayed response. It is a challenging task to design a robust classification system under above constraints with high accuracy. In this Letter, to resolve this problem, a novel architecture for weightless neural networks (WNNs) has been proposed. It uses variable sized random access memories to optimise the memory usage and a modified binary TRIE data structure for reducing the test time. In addition, a bio-inspired-based genetic algorithm has been employed to improve the accuracy. The proposed architecture is experimented on various disease datasets using its software and hardware realisations. The experimental results prove that the proposed architecture achieves better performance in terms of accuracy, memory saving and test time as compared to standard WNNs. It also outperforms in terms of accuracy as compared to conventional neural network-based classifiers. The proposed architecture is a powerful part of most of the low-power wearable devices for the solution of memory, accuracy and time issues.

Statistical speech reconstruction for larynx-related dysphonia has achieved good performance using Gaussian mixture models and, more recently, restricted Boltzmann machine arrays; however, deep neural network (DNN)-based systems have been hampered by the limited amount of training data available from individual voice-loss patients. The authors propose a novel DNN structure that allows a partially supervised training approach on spectral features from smaller data sets, yielding very good results compared with the current state-of-the-art.

Noise in ECG signals will affect the result of post-processing if left untreated. Since ECG is highly subjective, the linear denoising method with a specific threshold working well on one subject could fail on another. Therefore, in this Letter, sparse-based method, which represents every segment of signal using different linear combinations of atoms from a dictionary, is used to denoise ECG signals, with a view to myoelectric interference existing in ECG signals. Firstly, a denoising model for ECG signals is constructed. Then the model is solved by matching pursuit algorithm. In order to get better results, four kinds of dictionaries are investigated with the ECG signals from MIT-BIH arrhythmia database, compared with wavelet transform (WT)-based method. Signal–noise ratio (SNR) and mean square error (MSE) between estimated signal and original signal are used as indicators to evaluate the performance. The results show that by using the present method, the SNR is higher while the MSE between estimated signal and original signal is smaller.

Wearable Electrocardiogram (ECG) is attracting much attention in daily healthcare applications. From the viewpoint of long-term use, it is desired that the electrodes are non-contact with the human body. In this study, the authors propose an algorithm using the stationary wavelet transform (SWT) to remove motion artefact superimposed on ECG signal when using non-contact capacitively coupling electrodes. The authors evaluate the effect on motion artefact removal of this algorithm by applying it to various ECG signals with motion artefacts superimposed. As a result, the correlation coefficients of ECG signals with respect to the clean ones have been improved from 0.71 to 0.88 on median before and after motion artefact removal, which demonstrates the validity of the proposed SWT-based algorithm.

Real-time monitoring systems (RTMSs) have drawn considerable attentions in the last decade. Several commercial versions of RTMS for patient monitoring are available which are used by health care professionals. Though they are working satisfactorily on various communication protocols, their range, power consumption, data rate and cost are really bothered. In this study, the authors present an efficient embedded system based wireless health care monitoring system using ZigBee. Their system has a capability to transmit the data between two embedded systems through two transceivers over a long range. In this, wireless transmission has been applied through two categories. The first part which contains Arduino with ZigBee will send the signals to the second device, which contains Raspberry with ZigBee. The second device will measure the patient data and send it to the first device through ZigBee transceiver. The designed system is demonstrated on volunteers to measure the body temperature which is clinically important to monitor and diagnose for fever in the patients.

The outcome for patients diagnosed with facial palsy has been shown to be linked to rehabilitation. Dense 3D morphable models have been shown within the computer vision to create accurate representations of human faces even from single 2D images. This has the potential to provide feedback to both the patient and medical expert dealing with the rehabilitation plan. It is proposed that a framework for the creation and measuring of patient facial movement consisting of a hybrid 2D facial landmark fitting technique which shows better accuracy in testing than current methods and 3D model fitting.