80 research outputs found
Non-Contact Sleep Monitoring
No full text"The road ahead for preventive medicine seems clear. It is the delivery
of high quality, personalised (as opposed to depersonalised) comprehensive
medical care to all." Burney, Steiger, and Georges (1964)
This world's population is ageing, and this is set to intensify over the next forty years.
This demographic shift will result in signi�cant economic and societal burdens (partic-
ularly on healthcare systems). The instantiation of a proactive, preventative approach
to delivering healthcare is long recognised, yet is still proving challenging. Recent work
has focussed on enabling older adults to age in place in their own homes. This may
be realised through the recent technological advancements of a�ordable healthcare sen-
sors and systems which continuously support independent living, particularly through
longitudinally monitoring deviations in behavioural and health metrics. Overall health
status is contingent on multiple factors including, but not limited to, physical health,
mental health, and social and emotional wellbeing; sleep is implicitly linked to each of
these factors.
This thesis focusses on the investigation and development of an unobtrusive sleep mon-
itoring system, particularly suited towards long-term placement in the homes of older
adults. The Under Mattress Bed Sensor (UMBS) is an unobstrusive, pressure sensing
grid designed to infer bed times and bed exits, and also for the detection of development
of bedsores. This work extends the capacity of this sensor. Speci�cally, the novel contri-
butions contained within this thesis focus on an in-depth review of the state-of-the-art
advances in sleep monitoring, and the development and validation of algorithms which
extract and quantify UMBS-derived sleep metrics.
Preliminary experimental and community deployments investigated the suitability of the
sensor for long-term monitoring. Rigorous experimental development re�ned algorithms
which extract respiration rate as well as motion metrics which outperform traditional
forms of ambulatory sleep monitoring. Spatial, temporal, statistical and spatiotemporal
features were derived from UMBS data as a means of describing movement during sleep.
These features were compared across experimental, domestic and clinical data sets, and
across multiple sleeping episodes. Lastly, the optimal classi�er (built using a combina-
tion of the UMBS-derived features) was shown to infer sleep/wake state accurately and
reliably across both younger and older cohorts.
Through long-term deployment, it is envisaged that the UMBS-derived features (in-
cluding spatial, temporal, statistical and spatiotemporal features, respiration rate, and
sleep/wake state) may be used to provide unobtrusive, continuous insights into over-
all health status, the progression of the symptoms of chronic conditions, and allow the
objective measurement of daily (sleep/wake) patterns and routines
Effektivität und Effizienz von Online-Hilfesystemen in deutschen Universitäts-OPACs
No full textToday, almost always Online Public Access Catalogues (OPACs) are the only way to access library collections, but practical experience shows that users have problems dealing with electronic catalogues. The author therefore recommends that any good OPAC needs an elaborated online help system. The present master thesis surveys implemented online help systems in German university OPACs and gives a critical discussion. Beyond the variety of help texts, the author also includes help systems like spell check, search filters for results and recommender. On one hand, the analysis is based on a function-related classification of those help systems and on the other, the introduction of the terms effectiveness and efficiency serve as parameters to determine the usefulness of a help system. As a result, the author presents basic principles for an efficient help system that are based on the findings of this master thesis
Non-contact under-mattress sleep monitoring
No full textSleep quality and duration are increasingly recognised as being important prognostic parameters in the assessment of an individual's health. However, reliable non-invasive long-term monitoring of sleep in a non-clinical setting remains a challenging problem. This paper describes the validation of a novel under mattress pressure sensing sleep monitoring modality that can be seamlessly integrated into existing home environments and provides a pervasive and distributed solution for monitoring long-term changes in sleep patterns and sleep disorders in adults. 410 minutes of concomitant Under Mattress Bed Sensor (UMBS) and strain gauge data were analysed from eight healthy adults lying passively. In this analysis, customised respirations rate detection algorithms yielded a mean difference of −0.12 breaths per five minutes and a mean percentage error (MPE) of 0.16% when the sensor was placed beneath the mattress. 1,491 minutes of UMBS and video data were recorded simultaneously from four participants in order to assess the movement detection efficacy of customised UMBS algorithms. These algorithms yielded accuracies, sensitivities and specificities of over 90% when compared to a video-based movement detection gold standard. A reduced data set (267 minutes) of wrist actigraphy, the gold standard ambulatory sleep monitor, was recorded. The UMBS was shown to outperform the movement detection ability of wrist actigraphy and has the added advantage of not requiring active subject participation.</p
Physiological Signal Estimation and Movement Detection Using an Under-Mattress Pressure Sensing Grid During Sleep
No full text
Setwise Comparison: Consistent, Scalable, Continuum Labels for Computer Vision
No full textA growing number of domains, including affect recognition and movement analysis, require a single, real number ground truth label capturing some property of a video clip. We term this the provision of continuum labels. Unfortunately, there is often an uncacceptable trade-off between label consistency and the efficiency of the labelling process with current tools. We present a novel interaction technique, ‘setwise’ comparison, which leverages the intrinsic human capability for consistent relative judgements and the TrueSkill algorithm to solve this problem. We describe SorTable, a system demonstrating this technique. We conducted a real-world study where clinicians labelled videos of patients with multiple sclerosis for the ASSESS MS computer vision system. In assessing the efficiency-consistency trade-off of setwise versus pairwise comparison, we demonstrated that not only is setwise comparison more efficient, but it also elicits more consistent labels. We further consider how our findings relate to the interactive machine learning literature
If We’re Playing by the Rules, it has to be a Good Game
No full textIn the 2015 book Rule Breakers, author Niamh Hourigan postulates that Ireland’s success and downfall was due to personal relationships being much more important than rules. She says that a “weak rules and strong relationships” tension permeates Irish life, and not only at the top. This seems to have led to Ireland having a reputation of being friendly and open, and an easy place in which to do business. It has also created a place where corruption has been far too easy and widespread as a result of the power and influence of relationships
- …
