BEGIN:VCALENDAR VERSION:2.0 PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4// BEGIN:VEVENT UID:20260525T165632EDT-4166aDAgSU@132.216.98.100 DTSTAMP:20260525T205632Z DESCRIPTION:\n \n \n \n Abstract\n\n Cardiovascular disease (CVD) remains a lead ing cause of death and disability worldwide\, underscoring the critical ne ed for early detection and continuous monitoring. Wearable technologies ha ve emerged as a promising solution\, offering non-invasive\, real-time ass essment of physiological signals in daily life. Among these\, seismocardio graphy (SCG)—a technique that captures chest wall vibrations from cardiac activity using low-cost accelerometers—has the potential to enable afforda ble\, comfortable\, and continuous monitoring. However\, several limitatio ns hinder its widespread adoption. SCG signals are highly susceptible to m otion artifacts during ambulatory activity\, complicating their use in rea l-world settings. Furthermore\, the scarcity of large\, annotated SCG data sets limits the development and generalization of machine learning models. Compared to conventional modalities\, SCG remains underutilized for extra cting detailed cardiac features or supporting clinical use. This thesis ai ms to address these limitations and advance the utility of SCG for wearabl e cardiac monitoring by improving signal robustness\, mitigating data scar city\, and demonstrating its functional relevance across diverse applicati ons.\n\n First\, we develop a novel motion artifact reduction algorithm to improve SCG signal quality in ambulatory conditions. We show that our algo rithm improved heart rate estimation accuracy during walking\, up to a −19 dB signal-to-noise ratio without electrocardiography (ECG). Our solution is directly applicable to SCG monitoring in daily life. Second\, we tackle data scarcity with a generative adversarial network to create synthetic\, individualized SCG heartbeats. We show that we can successfully replicate SCG signal morphology with tunable features and demonstrate its utility i n a lung volume classification task\, with synthetic data matching the acc uracy of real data\, and an augmentation approach increased accuracy by 3% . Third\, we develop a generative adversarial network that uses 6-axis vib rational cardiography (VCG) to reconstruct multiple cardiac waveforms – el ectrocardiography (ECG)\, impedance cardiography (ICG)\, photoplethysmogra phy (PPG)\, and non-invasive blood pressure (NIBP). We demonstrate that th e approach achieves strong morphological and temporal alignment between es timated and reference signals\, with median Pearson correlation coefficien ts of 0.808\, 0.907\, 0.833\, and 0.929 for ECG\, NIBP\, ICG\, and PPG\, r espectively. This result demonstrates the feasibility of using a single mo tion sensor to estimate rich\, multimodal cardiac information\, offering a simplified and scalable alternative to traditional multi-sensor systems. Finally\, we explore the use of synthetic SCG data to improve cross-domain pulmonary hypertension (PH) detection. We leverage generative models and introduce a dataset selection method to optimize the composition of synthe tic and real training data. Our approach improves the out-of-distribution PH detection performance\, increasing the area under the ROC curve from 0. 51 to 0.86. These results highlight the potential of generative SCG modell ing in a clinically relevant scenario with limited training data.\n\n Colle ctively\, these contributions advance seismocardiography as a viable and s calable modality for wearable cardiac monitoring. By improving signal robu stness\, addressing data scarcity\, enabling multimodal estimation\, and d emonstrating clinical relevance\, this work expands the role of SCG in bot h research and real-world applications.\n \n \n \n\n DTSTART:20250716T170000Z DTEND:20250716T190000Z LOCATION:McConnell Engineering Building\, CA\, QC\, Montreal\, H3A 0E9\, 34 80 rue University SUMMARY:PhD defence of James Skoric – Wearable health monitoring with seism ocardiography and generative modeling URL:/ece/channels/event/phd-defence-james-skoric-weara ble-health-monitoring-seismocardiography-and-generative-modeling-366008 END:VEVENT END:VCALENDAR