Innovative Projects Realized

Uncertainty Modeling and Quantification in Neural Network Image Denoisers

Image denoising is a fundamental process in most of computer vision systems, imaging systems, and photography productions. Recently, with the power of deep neural networks, image denoising has been pushed towards new boundaries. However, neural network image denoisers are constrained by the accuracy of the noise model used to train them. Training on a poor noise model results in poor generalization performance on real-world images. Moreover, it is still unclear how to quantitatively assess the output of a neural network image denoiser, especially in real-world cases where ground truth is unavailable. To this end, this project aims at pushing the state-of-the-art denoising performance through: (1) training neural network image denoisers on more accurate noise models derived form real images; and (2) using loss functions that models and quantifies the uncertainty of the network outputs, hence, we can have a confidence measure of the denoising results.

Decision Support System to Manage and Forecast Final Project Costs

Managing construction projects can be a challenging task as they are subject to change and comprised of numerous work packages. Understanding project outputs in the early phases of a project can assist project managers to make informed decisions and take timely corrective actions to deliver projects on-time and on-budget. In the applicant’s doctoral research project, a Markov model was developed to enhance the accuracy of project cost forecasting through the integration of historical performance data. While this previous research has improved forecasting accuracy of a project’s ongoing cost performance, it cannot provide information regarding the impact that actions designed to mitigate poor cost performance (e.g., time crashing, duration extension) will have on project outcomes.
The objective of the current proposal is to develop a simulation-based framework that can predict the impact of various potential actions taken in response to poor project forecasts and suggest optimal actions. Monte Carlo simulation will be added to the previously developed Markov model to simulate various decision-making scenarios that can be taken by project managers. Outcomes of these analyses will be ranked and optimal decisions selected. Results of this research can be used as a supplementary tool to aid practitioners in the decision-making process.

Developing solutions for safer harvesting techniques on steep terrain – Year two

The forest industry in British Columbia (BC) is facing increasingly difficult challenges regarding fibre supply. New winch-assist technology that enables fully mechanized ground-based forest harvesting on steep terrain has been increasingly used in BC since 2016. The new systems have improved safety and provide access to fibre that was previously uneconomic. New low-consumption small-size cable yarders have also received increasing interest in non-trafficable terrain. Both the winch-assist and the cable yarding systems have similar issues related to the use of tensioned wire ropes and natural anchors (stumps or trees).
Rope and anchor failures have serious safety risks for the operators and the forest industry as a whole. To date, only a few studies focused on cable tensile force analysis and suitability of trees or stumps as anchors. This research aims to study the behaviour of cable-supported systems and anchors under varying conditions and improve safe practices of the newly introduced technology in BC.

Developing solutions for safer harvesting techniques on steep terrain

The forest industry in British Columbia (BC) is facing increasingly difficult challenges regarding fibre supply. New winch-assist technology that enables fully mechanized ground-based forest harvesting on steep terrain has been increasingly used in BC since 2016. The new systems have improved safety and provide access to fibre that was previously uneconomic. New low-consumption small-size cable yarders have also received increasing interest in non-trafficable terrain. Both the winch-assist and the cable yarding systems have similar issues related to the use of tensioned wire ropes and natural anchors (stumps or trees).
Rope and anchor failures have serious safety risks for the operators and the forest industry as a whole. To date, only a few studies focused on cable tensile force analysis and suitability of trees or stumps as anchors. This research aims to study the behaviour of cable-supported systems and anchors under varying conditions and improve safe practices of the newly introduced technology in BC.

Implementation of a low barrier hydromorphone distribution program to prevent fatal overdoses – Year two

North America is experiencing an unprecedented opioid overdose epidemic driven by the proliferation of fentanyl and fentanyl-adulterated drugs. Based at the BC Centre on Substance Use (BCCSU) and in collaboration with the BC Centre for Disease Control (BCCDC), I will undertake an ethno-epidemiological study to evaluate the implementation, uptake, and effectiveness of a novel low-barrier hydromorphone distribution program via a secure automated medication dispensing system targeting individuals at high risk of fatal overdose. The study will also examine barriers and facilitators to program scale up. The BCCDC is responsible for the implementation and operation of the distribution systems and program, with the BCCSU conducting the external evaluation. Data collection will involve qualitative interviews with 60 program participants and 15 stakeholders, as well as ethnographic fieldwork. The Vancouver Area Network of Drug Users’ (VANDU) mandate is to improve the lives of people who use drugs (PWUD), and they benefit by playing a key role in an evaluation and knowledge translation for a program with the potential to directly benefit PWUD. VANDU members will be included on a community advisory committee, be employed as peer research assistants, and be involved in future policy development and knowledge translation.

Implementation of a low barrier hydromorphone distribution program to prevent fatal overdoses

North America is experiencing an unprecedented opioid overdose epidemic driven by the proliferation of fentanyl and fentanyl-adulterated drugs. Based at the BC Centre on Substance Use (BCCSU) and in collaboration with the BC Centre for Disease Control (BCCDC), I will undertake an ethno-epidemiological study to evaluate the implementation, uptake, and effectiveness of a novel low-barrier hydromorphone distribution program via a secure automated medication dispensing system targeting individuals at high risk of fatal overdose. The study will also examine barriers and facilitators to program scale up. The BCCDC is responsible for the implementation and operation of the distribution systems and program, with the BCCSU conducting the external evaluation. Data collection will involve qualitative interviews with 60 program participants and 15 stakeholders, as well as ethnographic fieldwork. The Vancouver Area Network of Drug Users’ (VANDU) mandate is to improve the lives of people who use drugs (PWUD), and they benefit by playing a key role in an evaluation and knowledge translation for a program with the potential to directly benefit PWUD. VANDU members will be included on a community advisory committee, be employed as peer research assistants, and be involved in future policy development and knowledge translation.

Time Series Clustering and Classification

Financial indicators of an individual firm may be in the form of time series, vectors, or even richer data, such as text or images. The purpose of this work is to explore and develop methods for dealing with such data, and in particular perform the clustering/classification of such data into similar groups. In the project the intern will develop the tools that will allow to determine whether a client should be issued a loan or not.

Modeling Dune Roughness in a Large Sand-bedded River – Year two

Dunes are common bed features in large sand-bedded rivers and are significant sources of in-channel roughness. Reliably estimating roughness is crucial for the prediction of flood flows, but roughness is crudely represented in hydraulic models widely used by environmental consultants and engineers. Roughness is treated as a calibration knob in these models, adjusted to force predicted water levels to match observations. This makes water level predictions during floods difficult, especially for flows beyond the range of previous observations. The proposed project will examine the physical basis for dune-related roughness in rivers. Previous bed surveys from the sand-bedded Fraser River will be compiled and dune characteristics (height, length, lee angle) will be extracted. Roughness from model calibrations will be compared to observed dune dimensions. Recently proposed relations for predicting dune dimensions, developed separately, will be refined using the field observations and validated with some new field observations. Modules for existing hydrodynamic and morphodynamic models will be developed that incorporate the new predictive relations for roughness. The project will reduce reliance on time-consuming roughness calibration exercises. TO BE CONT’D

Modeling Dune Roughness in a Large Sand-bedded River

Dunes are common bed features in large sand-bedded rivers and are significant sources of in-channel roughness. Reliably estimating roughness is crucial for the prediction of flood flows, but roughness is crudely represented in hydraulic models widely used by environmental consultants and engineers. Roughness is treated as a calibration knob in these models, adjusted to force predicted water levels to match observations. This makes water level predictions during floods difficult, especially for flows beyond the range of previous observations. The proposed project will examine the physical basis for dune-related roughness in rivers. Previous bed surveys from the sand-bedded Fraser River will be compiled and dune characteristics (height, length, lee angle) will be extracted. Roughness from model calibrations will be compared to observed dune dimensions. Recently proposed relations for predicting dune dimensions, developed separately, will be refined using the field observations and validated with some new field observations. Modules for existing hydrodynamic and morphodynamic models will be developed that incorporate the new predictive relations for roughness. The project will reduce reliance on time-consuming roughness calibration exercises. TO BE CONT’D

Fonctionnement de l’erablierc a la limite de son aire de distribution

Le stage propose dans Ie cadre dc cette demande s’inscrit dans un projet plus vaste mene

par Dominique Gravel, Christian Messier, Yves Bergeron et moi:meme, subventionnc par

Ie FQRNT pour la periode allant de Janvier 2011 a Janvier 2015. L’objectifprincipal de

ce projet est d’avoir une meilleure comprehension des phenomenes sous-jacents is la

migration des especes vegetales a travers des zones deja occupees, de caracteriser les

parametres essentiels de cette dynamique et d’identifier de quelle fa~on les changements

de conditions climatiques peuvent influencer ces migrations.

Les travaux effectues dans Ie cadre de ce stage serviront a developper des modeles

statistiques qui permettront d’approfondir les modeles utilises jusqu’a present dans un but

de prospection du devenir des forets existantes. Une fois I’ensemble du modele PPA

etalonne et teste pour nos parcelles eXpCrimentales, nous serons en mesure d’evaluer les

differents scenarios possibles quant au devenir des zones en bordures d’ erablieres et du

potentiel economique de celles-ci……

Développement de biocides à base d’extractibles végétaux pour le contrôle des maladies et de la germination lors de l’entreposage des pommes de terre.

Un des marchés d’intérêt pour les molécules biocides des extraits d’écorces est l’agriculture, et plus spécifiquement la production de pommes de terre. Suivant la récolte des pommes de terre, leur entreposage peut devenir problématique. Deux grands défis se présentent; 1) les maladies c’est-à-dire le développement de microorganismes causant la pourriture des tubercules et 2) la germination hâtive des pommes de terre occasionne de nombreuses pertes. L’industrie de la pomme de terre au Canada évalue les pertes en entreposage à 10% de la récolte, ce qui représente environ 75 millions $. Au Québec, on évalue les pertes à 66 030 tonnes/an. Les fongicides et antigerminatifs chimiques actuellement utilisés pour contrer ces problématiques peuvent être dommageable pour l’environnement et même pour la santé. Peu d’alternative naturelle à coût abordable sont actuellement disponibles sur le marché. En accord avec leur perspective de développement durable et de préservation de l’environnement, l’UQTR, Innofibre et les Producteurs de Pommes de terre du Québec désire s’investir dans le développement de produits biosourcés pour répondre aux besoins des agriculteurs…

Development of Chitosan-based Nanosparticles forLoading Non-water Soluble Medicine

Over 90% new drugs cannot be used in market or perform inefficiently because of their poor water-solubility. Delivery of non-water soluble medicine by using biodegradable carriers can solve the problem. This research work aims to develop N-trimethyl chitosan chloride (TMC)-based nanoparticles used as a drug carrier for delivering non-water soluble medicine to enhance the therapeutic efficiency. Two interns will be involved in this project. Collaborating with Ovensa Inc., we will produce TMC-based nanoparticles with suitable particle size and surface properties. The loading efficiency, and release kinetics of the medicine loaded in TMC-based nanoparticles will be thoroughly investigated. The successful result of this project will eventually establish a platform for enhancing the activities of the non-water soluble medicine.