Adrian D.C. Chan - Profile on Academia.edu (original) (raw)

Papers by Adrian D.C. Chan

Driver identification using vehicle acceleration and deceleration events from naturalistic driving of older drivers

Driving is a common task that involves skill and individual preferences that can differ between d... more Driving is a common task that involves skill and individual preferences that can differ between drivers. The unique driving behaviours can be beneficial for differentiating drivers of shared vehicles and identifying differences between older drivers with normal and declining driving abilities. This paper presents a method for identifying individual drivers based on motor vehicle acceleration and deceleration events from their natural driving behaviour. We provide a novel approach to driver identification based on classification using multiple in-vehicle sensor signals collected in naturalistic conditions with anonymized driving locations. The dataset consists of thousands of trips from a selection of 14 stable-health older drivers (70 years and older) from their first year of the Candrive research study. We trained separate multiclass linear discriminant analysis classifiers to recognize unique patterns in their acceleration and deceleration events to predict the identity of the driver out of a group of drivers. For five different drivers, the acceleration and deceleration events were used to distinguish between drivers at 34% and 30% average accuracy, respectively. By taking a majority vote among the events, the accuracy improved to 61%, exceeding by about three times the null model of random guessing. This performance improvement continues when expanding the group from 2 to 14 drivers. The analysis shows potential for identifying drivers by the patterns in their driving maneuvers such as turning and stopping.

Journal of Biomechanics, Dec 1, 2019

This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

A Longitudinal Evaluation of the Impact of a Graduate Student Accessibility Training on Student Learning Outcomes

A Toolkit for Motion Artifact Signal Generation

2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA), Jun 22, 2022

The Paxon : An Electro-physical Model of a Myelinated Exon (poster)

CMBES Proceedings, May 21, 2019

Inertial measurement units (IMUs) have potential to be integrated into clinical assessments of mo... more Inertial measurement units (IMUs) have potential to be integrated into clinical assessments of movement-related disorders of the spine. This study evaluated 2 Mbientlab Meta-MotionR IMUs relative to Vicon motion capture equipment in tracking 3D spine motion during 35 cycles of constrained repetitive spine flexion-extension (FE) in 10 participants. Root-meansquare error (RMSE) was low in all anatomical planes (RMSE ≤ 2.43°). Pearson's correlation coefficient was strong in the FE and lateral bend (LB) planes (R ≥ 0.746), and weak-to-moderate in the axial twist (AT) plane (0.343 ≤ R ≤ 0.679). Additionally, there was very strong correlation between range of motion measurements in the FE plane (ICC2,1 = 0.99), and a wide range from weak to strong in the LB and AT planes (0.239 ≤ ICC2,1 ≤ 0.980). This study reveals that the IMUs perform well in tracking motion in the primary movement plane, and can be used for planar assessments of movement quality.

Signal Quality Assessment of Compressively Sensed Electrocardiogram

IEEE Transactions on Biomedical Engineering, Nov 1, 2022

Objective Develop a signal quality index (SQI) to determine the quality of compressively sensed e... more Objective Develop a signal quality index (SQI) to determine the quality of compressively sensed electrocardiogram (ECG) by estimating the signal-to-noise ratio (SNR). Methods The SQI used random forests, with the ratio of the standard deviations of an ECG segment and a clean ECG, and the Wasserstein metric between the amplitude distributions of an ECG segment and a clean ECG, as features. The SQI was tested using the Long-Term Atrial Fibrillation Database (LTAFDB) and the PhysioNet/CinC Challenge 2011 Database Set A (CinCDB). Clean ECG segments from the LTAFDB were corrupted using simulated motion artifact, with preset SNR between -12 dB and 12 dB. The CinCDB was used as-it-is. The databases were compressively sensed using three types of sensing matrices at three compression ratios (50%, 75%, and 95%). For LTAFDB, the RMSE and Spearman correlation between the SQI and the preset SNR were used for evaluation, while for CinCDB, accuracy and F1 score were used. Results The average RMSE was 3.18 dB and 3.47 dB in normal and abnormal ECG, respectively. The average Spearman correlation was 0.94 and 0.93 in normal and abnormal ECG, respectively. The average accuracy and F1 score were 0.90 and 0.88, respectively. Conclusion The SQI determined the quality of compressively sensed ECG and generalized across different databases. There was no consequential effect on the SQI due to abnormal ECG or compression using different sensing matrices and different compression ratios. Significance Without reconstruction, the SQI can inform which ECG should be analyzed to reduce false alarms due to contamination.

Non-obtrusive electrocardiogram system for the Smart Rollator

The objective of the Smart Rollator project is to enable health care monitoring through embedding... more The objective of the Smart Rollator project is to enable health care monitoring through embedding electronic sensing systems. This paper examines adding an ECG system to a rollator, designed specifically to interface with two gel-less electrodes. The focus is to develop a circuit with minimal filtering of the signal, just sufficient to extract valuable ECG information reliably using digital processing techniques. A prototype system is built and evaluated, its performance compared to a conventional three-electrode ECG amplifier. Preliminary results demonstrate that the proposed circuit can be used for the Smart Rollator application.

Sources of error during inertial sensing of human movement: a critical review of the fundamentals

Spectral Analysis of Respiratory and Cardiac Signals Using Doppler Radar

Springer eBooks, 2015

Inmate injuries and deaths remain a significant problem for correctional institutions, increasing... more Inmate injuries and deaths remain a significant problem for correctional institutions, increasing the need for continuous monitoring of inmates. A Doppler radar device is investigated for use as a contactless method of vital sign monitoring (e.g., breathing and heart rate) in a single cell setting. The recorded radar signal is analysed in both the time domain, and in the frequency domain. The radar signal and its frequency spectrum is compared against the signals and frequency spectrums obtained from an electrocardiogram and a respiratory inductance plethysmography band. The breathing and heart rate estimates obtained from the radar match up with the estimates provided by the respiratory band and electrocardiogram. Results show that the radar device demonstrates good potential for contactless vital sign monitoring.

Gating of false alarms in myocardial ischemia monitoring using ST segment deviation trend estimator

A false alarm gating system for myocardial ischemia monitoring is proposed to mitigate false alar... more A false alarm gating system for myocardial ischemia monitoring is proposed to mitigate false alarms resulting from inaccurate estimates of the ST deviation in the electrocardiogram (ECG). The proposed system employs multiple estimates of the ST segment deviation and correlates the trends between these estimates; low correlation can be indicative of an inaccurate estimate. Three correlation methods were considered: 1) Pearson correlation coefficient, 2) Kendall rank correlation, and 3) Spearman rank correlation. The proposed system was tested using 16 ECG signals from the Long-Term ST Database available on Physionet. The baseline performance of the commercial bedside monitor was 78 true alarms and 66 false alarms with precision and recall of 0.54 and 0.79, respectively. Using Spearman rank correlation, the proposed system balanced the gating of false alarms while minimizing the loss of true alarms. The system's true and false alarm rates were 71 and 30, respectively, while attaining precision and recall of 0.70 and 0.72, respectively.

Transfer Learning for Detection of Atrial Fibrillation in Deterministic Compressive Sensed ECG

Atrial Fibrillation (AF) is a cardiac condition resulting from uncoordinated contraction of the a... more Atrial Fibrillation (AF) is a cardiac condition resulting from uncoordinated contraction of the atria which may lead to an increase in the risk of heart attacks, strokes, and death. AF symptoms may go undetected and may require longterm monitoring of electrocardiogram (ECG) to be detected. Long-term ECG monitoring can generate a large amount of data which can increase power, storage, and the wireless transmission bandwidth of monitoring devices. Compressive Sensing (CS) is compression technique at the sampling stage which may save power, storage, and wireless bandwidth of monitoring devices. The reconstruction of compressive sensed ECG is a computationally expensive operation; therefore, detection of AF in compressive sensed ECG is warranted. This paper presents preliminary results of using deep learning to detect AF in deterministic compressive sensed ECG. MobileNetV2 convolutional neural network (CNN) was used in this paper. Transfer learning was utilized to leverage a pre-trained CNN with the final two layers retrained using 24 records from the Long-Term Atrial Fibrillation Database. The Short-Term Fourier Transform was used to generate spectrograms that were fed to the CNN. The CNN was tested on the MIT-BIH Atrial Fibrillation Database at the uncompressed, 50%, 75%, and 95% compressed ECG. The performance of the CNN was evaluated using weighted average precision (AP) and area under the curve (AUC) of the receiver operator curve (ROC). The CNN had AP of 0.80, 0.70, 0.70, and 0.57 at uncompressed, 50%, 75%, and 95% compression levels. The AUC was 0.87, 0.78, 0.79, and 0.75 at each compression level. The preliminary results show promise for using deep learning to detect AF in compressive sensed ECG.Clinical Relevance—This paper confirms that AF can be detected in compressive sensed ECG using deep learning, This will facilitate long-term ECG monitoring using wearable devices and will reduce adverse complications resulting from undiagnosed AF.

An ensemble of U-Net architecture variants for left atrial segmentation

Segmentation of the left atrium and proximal pulmonary veins is an important clinical step for di... more Segmentation of the left atrium and proximal pulmonary veins is an important clinical step for diagnosis of atrial fibrillation. However, the automatic segmentation of the left atrium from late gadolinium-enhanced magnetic resonance (LGE-MRI) images remains a challenging task due to differences in acquisition and large variability between individuals. Deep learning has shown to outperform traditional methodologies for segmentation in numerous tasks. A popular deep learning architecture for segmentation is the U-Net, which has shown promising results biomedical segmentation problems. Many newer network architectures have been proposed that leverage the base U-Net architecture such as attention U-Net, dense U-Net and residual U-Net. These models incorporate updated encoder blocks into the U-Net architecture to incrementally improve performance over the base U-Net. Currently, there is no comprehensive evaluation of performance between these models. In this study we (1) explore approaches for the segmentation of the left atrium based on different- Net architectures. (2) We compare and evaluate these on the STACOM 2018 Atrial Segmentation Challenge dataset and (3) ensemble these models to improve overall segmentation by reducing the internal variance between models and architectures. (4) Lastly, we define and build upon a U-Net framework to simplify development of novel U-Net inspired architectures. Our ensemble achieves a mean Dice similarity coefficient (DSC) of 92.1 ± 2.0% on a test set of twenty 3D LGE-MRI images, outperforming other fully automatic segmentation methodologies.

A Review of Techniques for Surface Electromyography Signal Quality Analysis

IEEE Reviews in Biomedical Engineering, 2023

Electromyography (EMG) signals are instrumental in a variety of applications including prosthetic... more Electromyography (EMG) signals are instrumental in a variety of applications including prosthetic control, muscle health assessment, rehabilitation, and workplace monitoring. Signal contaminants including noise, interference, and artifacts can degrade the quality of the EMG signal, leading to misinterpretation; therefore it is important to ensure that collected EMG signals are of sufficient quality prior to further analysis. A literature search was conducted to identify current approaches for detecting, identifying, and quantifying contaminants within surface EMG signals. We identified two main strategies: 1) bottom-up approaches for identifying specific and well-characterized contaminants and 2) top-down approaches for detecting anomalous EMG signals or outlier channels in high-density EMG arrays. The best type(s) of approaches are dependent on the circumstances of data collection including the environment, the susceptibility of the application to contaminants, and the resilience of the application to contaminants. Further research is needed for assessing EMG with multiple simultaneous contaminants, identifying ground-truths for clean EMG data, and developing user-friendly and autonomous methods for EMG signal quality analysis.

A Simple, LowCost, 3DScanning System Usingthe LaserLight-Sectioning Method

Theuseof3D scanning systems foracquiring the external shape features ofarbitrary objects hasmanya... more Theuseof3D scanning systems foracquiring the external shape features ofarbitrary objects hasmanyapplications inindustry, computer graphics, andmorerecently, thebiomedical field Thepotential exists toexpand theuseof3Dmodels even further, bycontinuing todevelop simpler, morecosteffective systems. Asimple, lost cost, 3Dscanning system ispresented which employs alaser light-sectioning technique. Results ofaproof of concept experiment fortheproposed system demonstrate thevalidity ofthechosen approach. Directions forfuture workarealso discussed.

For many electromyography (EMG) applications, a suitable dry electrode would simplify practical i... more For many electromyography (EMG) applications, a suitable dry electrode would simplify practical implementation of an EMG recording system. Wearable mobility monitoring is an example of such a system. Therefore, surface EMG signals, from Fraunhofer Institute for Biomedical Engineering (IBMT) flexible dry electrodes and Orbital Research electrodes, were compared to signals from conventional Ag/AgCl electrodes. EMG measurements were performed on the right tibialis anterior for a range of different activities, such as light twitches, isometric contractions, jumping, and walking. Signal feature comparisons, skin preparation effects (i.e., cleaning with isopropyl alcohol), and impedance-noise analyses were performed. Results showed that both dry electrodes had comparable sensitivity to the standard Ag/AgCl electrodes for detecting small unloaded muscle contractions and large loaded contractions. Results also showed that noise content and impedance are weakly correlated and skin preparation methods did not have a positive effect on skin/electrode impedance.

Evaluation of wearable IMU performance for orientation estimation and motion tracking

Introducing objective wearable IMU measurements of functional movement quality into clinical asse... more Introducing objective wearable IMU measurements of functional movement quality into clinical assessments may improve accuracy of diagnosis. The goal of the present study was to assess the performance of inexpensive wearable IMUs relative to conventional motion capture equipment during controlled movements that are representative of typical human movement. Thirty-five cycles of spine flexion-extension, lateral bending, and axial twisting were simulated by means of a motorized gimbal at speeds of 20 cycles/min and 40 cycles/min. Differences between cycle-to-cycle maximum angle, minimum angle, and ROM values, as well as correlational analyses within IMUs and between IMUs and motion capture, in all movement directions, were compared. All absolute differences in measurements were < 1.55°. There were very high correlations between repeated IMU measures (R > 0.99) in all movement directions showing reliability between sensors and measurements. Overall, it was revealed that the sensors perform very well in the primary movement direction and one secondary axis; however, correlation in the third axis is suboptimal for orientation estimation and motion tracking.

IEEE Transactions on Instrumentation and Measurement, Aug 1, 2018

Objective: This paper investigates the effect of applied pressure on the skin-electrode impedance... more Objective: This paper investigates the effect of applied pressure on the skin-electrode impedance. Applied pressure, which affects the skin-electrode impedance, can fluctuate in many acquisition setups, particularly in wearable devices. The skin-electrode impedance, in turn, impacts the quality of the recorded signal in biomedical monitoring devices. Methods: Three types of electrodes were examined: Ag/AgCl electrodes, conductive textile electrodes, and dry electrodes with surface microfeatures (Orbital Research Inc.). Impedance measurements were conducted as pressure was repeatedly applied (P = 4 kPa) and removed (P = 0 kPa) over several trials. A Cole-Cole impedance model was utilized to model the skin-electrode interface. Significance and Results: Results demonstrated large decreases in the skin-electrode impedance of dry electrodes (conductive textile and orbital electrodes), especially with the initial application of the pressure. Model parameters also proved to be highly dependent on the level of pressure in dry electrodes but less dependent and more stable in wet electrodes. Decreases in skin-electrode impedance associated with applied pressure were thought to be caused by an increased effective electrode contact area. Changes in skin-electrode impedance were irreversible, lasting even after the applied pressure was released. Differences skin-electrode impedance associated with changes in applied pressure, decreased as the number of trials increased. Conclusion: Applied pressure has larger effect on dry electrodes than wet electrodes. Wearable devices that employ dry electrodes may have poorer biomedical signal quality when initially donned; however, the advantage of wet electrodes with their lower sensitivity to applied pressure is diminished in long-term monitoring applications.

Screening Technologies to Support Infection Prevention and Control in Long-Term Care

Proceedings of the International Symposium of Human Factors and Ergonomics in Healthcare, Mar 1, 2023

Many clinical environments implemented Coronavirus Disease 2019 (COVID-19) screening technologies... more Many clinical environments implemented Coronavirus Disease 2019 (COVID-19) screening technologies at entry points to identify and isolate cases. Despite the widespread adoption of screening technologies in vulnerable care settings, such as in long-term care (LTC) homes, implementations and protocols have yet to be thoroughly reviewed in the literature. Therefore, the Dynamic Sustainability Framework was applied to identify the limitations of various screening technology implementations. Based on identified limitations, design recommendations are put forward to support healthcare planning and design teams of long-term care homes, and clinical facilities to improve infection prevention and control (IPC).

Frontiers in electronics, Jul 1, 2022

The objective of this paper is to develop an optimized system to detect Atrial Fibrillation (AF) ... more The objective of this paper is to develop an optimized system to detect Atrial Fibrillation (AF) in compressively sensed electrocardiogram (ECG) for long-term remote patient monitoring. A three-stage system was developed to 1) reject ECG of poor signal quality, 2) detect AF in compressively sensed ECG, and 3) detect AF in selectively reconstructed ECG. The Long-Term AF Database (LTAFDB), sampled at 128 Hz using a 12-bit ADC with a range of 20 mV, was used to validate the system. The LTAFDB had 83,315 normal and 82,435 AF rhythm 30 s ECG segments. Clean ECG from the LTAFDB was artificially contaminated with motion artifact to achieve −12 to 12 dB Signal-to-Noise Ratio (SNR) in steps of 3 dB. The contaminated ECG was compressively sensed at 50% and 75% compression ratio (CR). The system was evaluated using average precision (AP), the area under the curve (AUC) of the receiver operator characteristic curve, and the F1 score. The system was optimized to maximize the AP and minimize ECG rejection and reconstruction ratios. The optimized system for 50% CR had 0.72 AP, 0.63 AUC, and 0.58 F1 score, 0.38 rejection ratio, and 0.38 reconstruction ratio. The optimized system for 75% CR had 0.72 AP, 0.63 AUC, and 0.59 F1 score, 0.40 rejection ratio, and 0.35 reconstruction ratio. Challenges for longterm AF monitoring are the short battery life of monitors and the high false alarm rate due to artifacts. The proposed system improves the short battery life through compressive sensing while reducing false alarms (high AP) and ECG reconstruction (low reconstruction ratio).