14277 research outputs found
Sort by
Signal fusion and dimensionality reduction for classification and anomaly detection tasks
No full textThis thesis explores sensor fusion for feature reduction applied to classification
and anomaly detection tasks. We developed new signal fusion techniques applied to
tactile sensor signals texture classification and dimensionality reduction for anomaly
detection in time series data. The first part of this thesis introduces a novel approach
dimensionality reduction of tactile signals for texture classification using principal
component analysis (PCA) and reducing exploration time without compromising
classifier accuracy. Various pipeline configurations demonstrated that a 3-second
exploration combined with PCA-fused features achieved up to 98% classification accuracy
with a significant reduction in feature input, yielding a reduction factor of
6750 times. This part of the work highlights PCA's advantage over alternative fusion
techniques, offering both interpretability and dimensionality reduction that enhance
classifier performance.
The second part of this thesis examines anomaly detection within thread line
signals, comparing the efficacy of PCA-based fusion, averaging, and raw signal methods.
Experimental results across three thread lines indicate that PCA-based fusion
provides a balanced sensitivity to anomalies, offering a streamlined process by removing
the need for individual signal threshold adjustments. This work contributes to
advancements in classification and automated anomaly detection by showcasing the
effectiveness of PCA-based signal fusion for feature reduction and robust anomaly
detection.Includes bibliographical references (pages 66-74
The effect of neonatal total parenteral nutrition on glucose metabolism in neonates and adult Yucatan miniature pigs
No full textTotal parenteral nutrition (TPN) is used when oral nutrition is not possible, but may cause
metabolic disturbances, increasing the risk of Type 2 diabetes mellitus (T2DM). We
hypothesize that TPN feeding early in life can alter glucose metabolism in a way that
persists into adulthood, leading to the development of biomarkers associated with T2D.
Additionally, we hypothesized that supplementing TPN with betaine and creatine could
potentially correct these changes, and that intrauterine growth-restriction (IUGR) could
exacerbate TPN-induced changes. We assigned 32 female Yucatan miniature piglets to four
groups: normal birth weight receiving TPN (TPN); sow-fed (SF); normal birthweight TPN
supplemented with betaine and creatine (TPN-B+C); and IUGR piglets fed TPN (TPN�IUGR). After 2 weeks on TPN (or SF), glucose metabolism and insulin sensitivity was
assessed. All pigs were then fed an oral diet for ~10 mo, and glucose metabolism tests were
repeated. TPN feeding showed significantly more sensitive glucose metabolism, which
were more pronounced immediately after TPN but remained significant 10 mo later. TPN
also increased insulin sensitivity, which was corrected by adding betaine and creatine.
IUGR did not exacerbate TPN effects. This study suggests that TPN in early life can
permanently impact glucose metabolism into adulthood, but these changes do not align
with T2DM.Includes bibliographical references (pages 74-90
Removal and separation of lead, copper and cadmium from water solutions using resin adsorption
No full textHeavy metals including lead, copper, and cadmium existing in various industrial effluents create severe environmental impacts around the world. There are various technologies applied for the treatment of wastewaters bearing toxic metals such as ultrafiltration, solvent extraction, precipitation, reverse osmosis, and resin adsorption via ion exchange or chelating bonding. Removal and separation of heavy metals have been the subject of many studies, but few of them systematically investigated various resins at the same time and the acting mechanism of functional group types in details. The current research bridged this gap with the following novel works : (1) experimental investigation and molecular simulation of uptake of cadmium ions from aqueous solution using resins with sulphonic /phosphonic functional groups; (2) removal/ recovery of single copper and lead ions from water system with comprehensive studies of the influences of pH, adsorption time, resin dosage, concentrations, as well as adsorption kinetics, thermodynamics and resin elution process; (3) experimental investigation of the separation of lead and copper ions from acidic binary solution, systematic studies on various operational conditions at different measurement ranges, as well as kinetics, isotherms, thermodynamics and elution studies. The results from the research demonstrated the potency for removal of lead, copper, and cadmium as well as lead-copper separation from wastewaters by the use of resin adsorption via ion exchange or chelating bonding. More resin structure-performance studies will be conducted in the future for efficient metal removal/recovery from wastewaters, secondary resources, or even ocean waters.Includes bibliographical reference
The development of an educational resource on mouth care and oral mucositis for patients receiving chemotherapy in Newfoundland and Labrador
No full textBackground: Oral mucositis is a common side effect of chemotherapy. The development of oral
mucositis has many negative impacts on the patient receiving chemotherapy and the healthcare
system, such as reduced oral intake, increased risk for infection, increased need for medical
intervention, interruption to chemotherapy, and increased healthcare costs. Prevention measures
and early intervention should be considered to reduce the impact of oral mucositis. Patient
education can be used to educate patients receiving chemotherapy about oral mucositis,
prevention, early identification and intervention to reduce the impact of oral mucositis. Purpose:
This practicum project aimed to develop an educational resource on mouth care and oral
mucositis for patients receiving chemotherapy in Newfoundland and Labrador. Methods: Three
methods were used to gather information on mouth care and oral mucositis in patients receiving
chemotherapy and explore education as an intervention to address this clinical issue. The three
methods used included an integrative literature review, an environmental scan, and consultations
with key stakeholders. Results: An educational resource, including a resource manual and patient
pamphlet, was developed based on the methods used. Orem’s Self-Care Deficit Nursing Theory
was used to guide the development of the educational resource. Conclusion: An educational
resource on mouth care and oral mucositis was developed for patients receiving chemotherapy in
Newfoundland and Labrador. There was no implementation or evaluation component for this
practicum project due to time restraints of the course. A plan for future evaluation of the
resources will be outlined in this report.Includes bibliographical references (pages 97-100
Implementation of novel deep learning algorithms for the high-resolution study of the lacuno-canalicular network from individuals with a documented history of chronic opioid use
No full textBone is a dynamic tissue that changes throughout life. This process is governed by
osteocytes that exist in a lacuno-canalicular network (LCN), but is altered by several factors,
including exercise, age, nutrition, and substance use. Artificial intelligence brought several
enhancements to image segmentation for medical imaging. However, it has not been applied to
study the LCN in human bone. This thesis implements novel deep learning methods on
Synchrotron Radiation micro-Computed Tomography (SRμCT) datasets of human rib cortical
bone microstructure to characterize osteoporosis-related features.
Ninety-seven human left sixth rib specimens (male: n = 60, female n = 37) were excised
from cadavers with informed consent. The specimens were divided into age categories defined by
decade. A 50-slice subset from six samples was segmented to train the U-Net++ deep learning
model. It was compared to traditional and manual segmentation methods. Deep learning performed
comparably to the traditional method, although it was more time-efficient. A follow-up model with
the MA-Net architecture more accurately segmented the data. Comparing segmented
microstructural parameters with opioid use, sex, and age revealed age as the most significant
predictor of deteriorating bone health. The results did not provide strong evidence of drug-induced
impacts on bone health as originally predicted, however, there are some indications hinting at a
link between opioid use and bone health. A follow-up study implementing a rabbit model is
underway to eliminate confounding factors present in a human population. However, this project
successfully created a novel segmentation algorithm that performed more efficiently in SRμCT
data segmentation.Includes bibliographical references (pages 111-131
Development of a virtual peritoneal dialysis resource for registered nurses
No full textBackground: Peritoneal dialysis (PD) is a common renal replacement therapy for patients with
end-stage renal disease (ESRD) and chronic kidney disease (CKD). Poor care of patients who
use PD could lead to severe complications, including peritonitis, exit-site infection, technique
failure and death. To ensure safe and effective peritoneal dialysis, nurses must receive
comprehensive education and training on the proper techniques and procedures involved in the
process. This will help them develop the skills and expertise to perform PD tasks proficiently and
provide the best care to patients undergoing dialysis treatment. Purpose: To assess current
evidence on virtual education methods to develop an educational resource for registered nurses
providing care to patients who use peritoneal dialysis. Methods: Three methods were used to
collect information for this project. First, an integrative literature review was conducted using a
literature search in CINAHL, PubMed, and Scopus. Eleven relevant articles were identified, and
research studies were critically appraised using standardized appraisal tools. Second,
consultations with stakeholders (i.e., registered nurses, physicians, managers, patient care
facilitators, and vendor representatives) were conducted to determine the learning needs of
registered nurses in Newfoundland and Labrador (NL) and gather recommendations and
feedback from all stakeholders regarding the development of a virtual peritoneal dialysis
resource. Finally, an environmental scan was conducted to determine current resources and the
best options for registered nurses. Findings: Registered nurses were effectively educated on
peritoneal dialysis through virtual education methods, which included computer-assisted
programs, e-modules, and Microsoft Teams. These findings aligned with the outcomes of the
consultations and environmental scan. Conclusion: Limited literature exists on virtual methods
to educate nurses about peritoneal dialysis. However, available research confirms the
effectiveness of online methods for this purpose. An evidence-based virtual peritoneal dialysis
resource has been developed to aid nurses in caring for patients who use peritoneal dialysis. This
virtual resource will be shared with Newfoundland and Labrador Health Services (NLHS)
members so they can plan to implement it.Includes bibliographical references on page 16
Adaptations in physiological and neuronal function during diet-induced obesity
No full textObesity significantly increases the risk of developing chronic conditions including type II diabetes, cardiovascular disease, and some cancers. The rate of obesity has tripled globally since 1975, which is in part due to the sudden prevalence and overconsumption of palatable high-fat diets (HFDs). Obesity profoundly perturbs the neural control of energy balance, affecting diverse cell types within the hypothalamus. However, an incomplete understanding of how HFD impacts the regulation of energy balance hinders our ability to more effectively treat obesity.
In this thesis, I describe the physiological and neuronal response to HFD feeding in rodents. We identified that HFD exposure elevates the body weight set point, which is initially driven by a transient hyperphagia. This hyperphagia coincides with increased excitatory transmission to lateral hypothalamic orexin (ORX) neurons, which regulate acute food intake. This suggests that ORX neurons may be involved in the initial hyperphagia, implicating them in the development of obesity. As HFD prolongs, body weight gain slows and reaches a new steady state regardless of age at the start, duration of feeding, or palatability of the diet. This sustained weight coincides with increased synaptic contacts to melanin-concentrating hormone (MCH) neurons, which promote weight gain and food intake, likely contributing to the maintenance of obesity.
The molecular mechanism underlying the establishment of a new set point remains elusive. During HFD feeding, the presence of a chronic low-grade
hypothalamic inflammation exacerbates weight gain, therefore we reasoned that inflammatory factors could modulate appetite-promoting neurons to maintain a new set point. We found that the inflammatory mediator prostaglandin E2 (PGE2) activate MCH neurons via its EP2 receptor (EP2R). Suppressing PGE2-EP2R on MCH neurons partially protects against excess weight gain and fat accumulation in the liver during HFD feeding. This mechanism could contribute to the maintenance of an elevated body weight set point in during diet-induced obesity.
Without long-term treatment options in face of the increasing rates of obesity, we are in desperate need of novel interventions. In the future, we hope that targeting EP2R on MCH neurons can lower body weight set point and aid in combatting obesity.Includes bibliographical references (pages 169-216
Business model innovation in start-ups: an exploratory case study of why and how business models are changed
No full textIt has been long understood that start-ups change their business models. However,
research on creating a business model, called business model development, and the change of
business model, called business model innovation, has primarily focused on established firms.
There is a lack of empirical evidence of why and how start-ups change their business model, and
it is unclear to what extent existing literature on established firms can be applied to start-ups. The
research question of this thesis is “Why and how do start-ups change their business models?”.
This thesis provides a unique contribution to the study of business model innovation by
providing an exploratory case study of six versions of a start-up’s business model canvas. While
considering the impact of human capital investments and outcomes, it is shown that a) business
models are changed because founders believe that the business model is not, or cannot be,
profitable, OR that there is a more profitable and scalable option within reach, and b) business
models are changed by discovering new markets or customer use cases and then arranging the
resources, capabilities, network allies, and operations necessary to achieve the desired impact.Includes bibliographical references (pages 76-82
Decadal changes in biomass and distribution of key fisheries species on Newfoundland’s Grand Banks
No full textCanadian fisheries management has embraced the precautionary approach and the incorporation
of ecosystem information into decision-making processes. Accurate estimation of
fish stock biomass is crucial for ensuring sustainable exploitation of marine resources. Spatio-
temporal models can provide improved indices of biomass as they capture spatial and
temporal correlations in data and can account for environmental factors influencing biomass
distributions. In this study, we developed a spatio-temporal generalized additive model (st-
GAM) to investigate the relationships between bottom temperature, depth, and the biomass
of three key fished species on The Grand Banks: snow crab (Chionoecetes opilio), yellowtail
flounder (Limanda ferruginea), and Atlantic cod (Gadus morhua). Our findings revealed
changes in the centre of gravity of Atlantic cod that could be related to a northern shift of the
species within the Grand Banks or to a faster recovery of the 2J3KL stock. Atlantic cod also
displayed hyperaggregation behaviour with the species showing a continuous distribution
over the Grand Banks when biomass is high. These findings suggest a joint stock assessment
between the 2J3KL and 3NO stocks would be advisable. However, barriers may need
to be addressed to achieve collaboration between the two distinct regulatory bodies (i.e.,
DFO and NAFO) in charge of managing the stocks. Snow crab and yellowtail flounder centres
of gravity have remained relatively constant over time. We also estimated novel indices
of biomass, informed by environmental factors. Our study represents a step towards ecosystem-
based fisheries management for the highly dynamic Grand Banks
ColocZStats: a 3D Slicer extension for assessing colocalization in confocal microscopy
No full textConfocal microscopy has evolved as a widely adopted imaging technique in molecular
biology and is frequently utilized to achieve accurate subcellular localization of
proteins. Applying colocalization analysis on image z-stacks obtained from confocal
fluorescence microscopes is a dependable method to reveal the association between
different molecules. In addition, despite the established advantages and growing adoption
of virtual reality (VR) technology in various microscopy research domains, there
has been a scarcity of systems supporting colocalization analysis within VR space.
In this context, several broadly employed biological image visualization platforms
were meticulously explored in this study to comprehend the current landscape. It
has been observed that while these applications can generate three-dimensional (3D)
reconstructions for the z-stacks and transfer them into an immersive VR scene, there is
still a common necessity for them to optimize the capability for executing quantitative
colocalization analysis on such images.
To constructively improve the above circumstances, an extension called ColocZStats
has been developed for 3D Slicer. With a user-friendly interface, ColocZStats
allows investigators to conduct intensity thresholding and region-of-interest (ROI)
selection on imported 3D image stacks. It can deliver several essential colocalization
metrics for structures of interest in the form of diagrams and spreadsheets. While
currently serving as a desktop tool, ColocZStats has been continuously enhanced and
will be systematically extended into VR in the next phase.Includes bibliographical references (pages 58-76