The University of Southampton

node type automatic_tidt_calendar_event has no template

Dr Peter Charlton from King’s College London will be visiting on Tuesday 24th April. He will give a talk entitled “Achieving clinical quality from wearable sensors: the role of signal processing”

The talk will take place in the Demo Room 59/4240 at 10:00am.

Abstract:

The development of wearable sensors such as smart watches provides a new opportunity to continuously track peoples’ health. In his talk, Peter Charlton will introduce some of the key challenges posed by wearable sensor data, and the solutions provided by signal processing techniques: signal quality assessment, extracting physiological measurements, and predicting clinical outcomes. Peter works in collaboration with clinicians at King’s College London to develop signal processing techniques to meet these challenges, and translate his research into clinical practice. His talk will draw on examples from both laboratory experiments and clinical studies.

Wearable sensor data are highly susceptible to artifact, which can be caused by motion and poor sensor contact. Signal quality assessment tools can be used to mitigate against artifact, either filtering the signal to eliminate artifact, or identifying high quality periods of signal from which reliable measurements can be obtained. Peter will present signal quality indicators designed for the electrocardiogram, pulse oximetry and respiratory signals.

Wearable sensor signals are influenced by a range of physiological systems, including the cardiac, vascular, autonomic and respiratory systems. Consequently, algorithms have been designed to extract a wide range of physiological measurements from the signals. Peter will provide an overview of signal processing techniques used to extract routine parameters such as heart rate and blood pressure. He will then describe his recent research into extracting indicators of respiratory and vascular state.

The final challenge of exploiting wearable sensor data is to present it to patients and clinicians in a concise and informative manner. Several approaches can be used, including eliminating outlying measurements, providing summary statistics, and generating intelligent alerts when a patient’s physiology changes for the worse. Peter will present a novel approach to analyse wearable sensor data to identify hospital patients at increased risk of clinical deteriorations such as heart attacks and strokes.

f83b1e2bb226e671ad5e39a6f6225aecB65Thursday, March 28, 2024 - 20:32https://www.sems.ecs.soton.ac.uk/events/B65ECS SeminarsECS Events<a style='color:white' href='http://data.southampton.ac.uk/building/59'>New Zepler (59)</a> - 4240http://data.southampton.ac.uk/building/59.map