Machine-Learning-Based Artistic Photo Manipulation and Stylization on Mobile Devices

Deep Learning is the new, and fast growing bread of systems based on artificial neural networks. These are phenomenal at learning patterns and generating new pattern that have the characteristics of the data the system was trained on. One of the main artistic and creative use of Deep Learning is to process images, pictures video. One typical task is that of style transfer by which the style of one image is used to shape the rendering of the one provided by the user.

Challenge for Change (C4C) Listening Projects

Building on Canada’s rich history as a participatory media producer, the project partners aim to use the democratic power and potential of media to help build a stronger, more inclusive Canada by creating opportunities for Canadians to talk and especially listen to each other: one-on-one, with friends, mothers, daughters, neighbours and strangers.

From trees to bluebirds: the communication of conservation on Vancouver Island

The Garry Oak Ecosystems Recovery Team (GOERT) is a non-government organization that endeavours to inform the public of their scientific findings and conservation goals. Social media has been used for conservation outreach, but the role of the public in sharing tweets and posts can have mixed effects. What information are shared by the public can either further share scientific information, maintain false information in the public sphere, or be ignored without any uptake.

Feasibility Study of a TV-Based Video Chat Technology for Elderly with Dementia

According to the Canadian Institute for Health Information, many of Canada's increasing aging population experience physical and cognitive impairments that necessitate moving to a care facility, with 44% of these residents having or indicating symptoms of depression. Social isolation is a contributor; 22-42% of these residents reported feeling severe loneliness. Video chat sessions have been found to be valuable augments to in-person visits. However, for adults with impairments such as dementia, operating the associated technologies might be impossible.

Software Framework for Smart Building Energy Audits

With more and more buildings being controlled by automation systems, one would expect their energy performance to be optimised. This is not the case however. Buildings can still go out of tune, and building operators can become overwhelmed by the alarms sounding from the automation systems, not knowing how to prioritize them. SES consulting is well poised to provide a “human in the loop” performance analysis service, leveraging their expert knowledge and the data from the building automation system.

Prediction of user intentions based on video captured interaction

User intention prediction of event data, collected during the business process, has become an important topic in Web Analysis and Business Intelligence. Commercial organizations have realized its importance for providing cost-effective opportunities to improve their decision-making in digital marketing strategies. We aim to develop and implement a statistical prediction model to make the prediction of user intention, using the retroactive video tracking data, while the anonymous customer navigates on the website.

Developing Prediction Models on London Stock Exchange (LSE) Equitiesand Indicies using Microsoft Azure Machine Learning and Data Mining

I am to import ten year’s worth of amassed historical data on news events, price movement of equities and public sentiment metrics to Microsoft Azure platform for study and analysis through the latest Data Mining techniques with an Economics point of view to uncover the hidden correlation and casualty between events and price movement of global markets in multiple timeframes (three hours, daily, weekly, monthly and yearly).

Image matching for purposes of consumer recommendation

The purpose of this project is to develop a highly accurate e-commerce recommender system able to select products across databases and recommend them to prospective customers both in real-time and off-line. Leveraging the historical inventory of sold products, browsing history, purchase history, and expressed preferences helps the recommender to formulate highly accurate product suggestions to find closest matches to what a consumer is looking for.

Rhetorical foundations of procedurally generated game worlds

The PC and console video game market have overtaken Hollywood movies as the number one source of entertainment dollars spent in North America. Very little work has been done to date on the effects that these games are having on (particularly young) players. This project will take well established rhetorical principles (Burke, Fahnestock, Perlman) that have been applied to the analysis of many other multi-media entertainment products and try to build a framework for a video game rhetoric.

Color management for ultra-bright conditions in OLED Displays

The objective of the proposed research project is to develop a realistic color appearance model based on the human visual system’s functioning that addresses the issue of reflections under high luminance levels. This will be incorporated into algorithms used in the color reproduction and retargeting algorithms of OLED displays. This model should give rise to reduced power consumption in OLED displays, while maintaining a high perceived quality of images. The project will also explore the effect of display reflections from OLED screens.