Search

This paper presents a new approach to gait analysis and parameter estimation from a single miniaturized ear-worn sensor embedded with a triaxial accelerometer. Singular spectrum analysis combined with the longest common subsequence algorithm has been used as a basis for gait parameter estimation. It incorporates information from all axes of the accelerometer to estimate parameters including swing, stance, and stride times. Rather than only using local features of the raw signals, the periodicity of the signals is also taken into account. The hypotheses tested by this study include: 1) how accurate is the ear-worn sensor in terms of gait parameter extraction compared to the use of an instrumented treadmill; 2) does the ear-worn sensor provide a feasible option for assessment and quantification of gait pattern changes. Key gait events for normal subjects such as heel contact and toe off are validated with a high-speed camera, as well as a force-plate instrumented treadmill. Ten healthy adults walked for 20 min on a treadmill with an increasing incline of 2% every 2 min. The upper and lower limits of the absolute errors using 95% confidence intervals for swing, stance, and stride times were obtained as \35.5 \pm \3.99 \ms, \36.9 \pm \3.84\ ms, and \17.9 \pm \2.29\ ms, respectively. [ABSTRACT FROM PUBLISHER]

The application of machine learning-based tele-rehabilitation faces the challenge of limited availability of data. To overcome this challenge, data augmentation techniques are commonly employed to generate synthetic data that reflect the configurations of real data. One such promising data augmentation technique is the Generative Adversarial Network (GAN). However, GANs have been found to suffer from mode collapse, a common issue where the generated data fails to capture all the relevant information from the original dataset. In this paper, we aim to address the problem of mode collapse in GAN-based data augmentation techniques for post-stroke assessment. We applied the GAN to generate synthetic data for two post-stroke rehabilitation datasets and observed that the original GAN suffered from mode collapse, as expected. To address this issue, we propose a Time Series Siamese GAN (TS-SGAN) that incorporates a Siamese network and an additional discriminator. Our analysis, using the longest common sub-sequence (LCSS), demonstrates that TS-SGAN generates data uniformly for all elements of two testing datasets, in contrast to the original GAN. To further evaluate the effectiveness of TS-SGAN, we encode the generated dataset into images using Gramian Angular Field and classify them using ResNet-18. Our results show that TS-SGAN achieves a significant accuracy increase of classification accuracy (35.2%-42.07%) for both selected datasets. This represents a substantial improvement over the original GAN.

In this paper, a new simple index has been introduced for the assessment of electrocardiography (ECG) signal quality. In the proposed method, first, the initial spectrum of the ECG is derived by applying synchrosqueezed wavelet transform (SSWT). Then, the main frequency rhythm of heart rate with maximum-energy embedded in the spectrum of the ECG signal is reconstructed using time-frequency ridge estimation algorithm. The ridge is subjected to the inverse SSW and SSW subsequently to reconstruct a clear spectrum corresponding to the main heart rhythm. Subtracting it from the initial spectrum, the resulting differential spectrum is converted to a single time-series by simply summing all the energy levels at each time-point. It has been shown that the derived time-series is proportional to the quality of ECG signal in terms of preserving its physiological features. The results of this research provide a profound basis for signal quality assessment of both ECG and photoplethysmography (PPG) signals under various noisy conditions and abnormal heart rate.

This paper presents a new method for estimating the average heart rate from a foot/ankle worn photoplethysmography (PPG) sensor during fast bike activity. Placing the PPG sensor on the lower half of the body allows more energy to be collected from energy harvesting in order to give a power autonomous sensor node, but comes at the cost of introducing significant motion interference into the PPG trace. We present a normalised least mean square adaptive filter and short-time Fourier transform based algorithm for estimating heart rate in the presence of this motion contamination. Results from 8 subjects show the new algorithm has an average error of 9 beats-per-minute when compared to an ECG gold standard.

A novel mathematical model based on multi-way data construction and analysis with the goal of simultaneously separating and localizing the brain sources specially the subcomponents of event related potentials (ERPs) is introduced. We represent multi-channel EEG data using a third-order tensor with modes: space (channels), time samples, and number of segments. Then, a multi-way technique, in particular, generalized version of PARAFAC2 method, is developed to blindly separate and localize mutually/temporally correlated P3a and P3b sources as subcomponents of P300 signal. In this paper the non-orthogonality of the ERP subcomponents is defined within the tensor model. In order to obtain essentially unique estimation of the signal components one parametric and one structural constraint are defined and imposed. The method is applied to both simulated and real data and has been shown to perform very well even in low signal to noise ratio situations. In addition, the method is compared with spatial principal component analysis (sPCA) and its superiority is demonstrated by using simulated signals.