Trend-Detector – A Fixation Detection Approach to Deriving Trends in Real-time ICU Monitor Data

A. Salatian (UK) and N. Whitehead (Iceland)


Trend detection, signal processing, eye tracking, medicine


The monitors in an Intensive Care Unit produce very high volumes of noisy continuous data at very high sampling rates for each physiological signal. This is compounded when numerous signals are being recorded simultaneously. It is computationally expensive to reason on a point to point basis. We believe it is better to convert the data into trends in real-time this would allow medical staff or a higher level process to reason qualitatively with trends which are increasing, decreasing or steady. In this paper we present TREND-DETECTOR, a novel dynamic approach to real-time trend detection using a fixation detection technique from the eye tracking community.

Important Links:

Go Back