Design of Linear-Phase Mirror-Image Symmetric Filter Banks for Image Compression Coding

The discrete cosine transform (DCT) has widely been used in image and video compression coding. However, the conventional DCT causes severe block artifacts in the compressed image in low-bit-rate coding applications. The block effect is perceived as visible discontinuities cross the block boundaries. Many researchers have proposed other techniques, such as the filter bank approach, to remove or suppress this undesirable feature.

Design and Implementation of IEEE 802.16j MMR Networks

Sparkmatrix, a provider of online tools for property managers, has positioned one of its next-generation service provisioning commodities at IEEE 802.16 networks (or Wimax). The company successfully delivered IEEE 802.11 and IEEE 802.16 services to their customers. To improve the service quality and coverage, they considered deploying relays in order to erase any possible black notes and service deficiencies in buildings and offices.

Assymetric Lossless Compression via Algorithmic Self-organizing Structures

This internship project with Deep Compression Inc. aims to produce a prototype demonstrating a new technique for asymmetric lossless data compression. Unlike conventional compression techniques, this prototype will allow representational information to be recovered using an algorithmic recovery procedure, algorithmic self-organizing structure (ASOS).

OpenVL Implementation on Multiprocessor Platform

The project involves creating a vision based system to perform multi-camera 3D tracking of persons in an indoor environment. The implementation of this system will be using Open Vision Library (a standard library for image processing that is designed with hardware acceleration in mind) and it will be running on one or multiple processor hardware platform developed by LightHaus. LightHaus Logic is a Vancouver-based company that develops compact high-performance systems for advanced video analytics and embedded computer applications.

Game Object Parallelization

In order to produce ever more powerful computer chips designers have begun to produce CPUs with duplicated infrastructure and machines with multiple CPUs. This allows these devices to perform actual simultaneous calculation. The most recent generation of video game consoles are built with this manner of technology, thus adding a new dimension of complexity to the design of the underlying system. Taking a traditional computational task and dividing it for parallel execution is not trivial.

Stochastic Modelling of Pre-sales and Network Request Provisioning Sub-processes at Bell Canada

The objective of this project is to contribute a quantitative framework for analyzing potential inefficiencies of current practices at Bell Canada in the early stages of telecommunication service provisioning, particularly pre-sales and network request sub-processes. Models of these processes incorporating randomness are to be estimated from past data and utilized as a basis for efficiency improvement using both simulation and analytical optimization tools.

Monitoring Sleep-related Breathing Irregularities with Computer Vision Techniques

Sleep monitoring plays an essential role in the diagnosis of sleep disturbances. While the polysomnogram represents the gold standard, it has the disadvantage of assessing the patient’s sleep in an unfamiliar environment, usually a sleep clinic. This proposal targets the development of new computer vision-based methods for home-based sleep monitoring in partnership with Vigil Health Solutions, a company providing healthcare services and products internationally.

Lie Algebra Image Processing Applied to Intrinsic Optical Imaging

Optical imaging and the underlying image analysis has seen tremendous progress in recent years. Although it is now possible to perform multi-modal acquisition, the analysis frameworks for multi-modal data remain elusive. For example, the different scales and resolutions at which the images are taken require the use of mathematical techniques to deform, analyze and co-register images in a coherent framework. In parallel, J. Patera and his research group have, over the last few years, developed new Lie Algebra-based techniques to process images and perform continuous operations on them.

Cryptographic Security Solution for Laptops

In their various versions (XP, 2003, Vista, Longhorn), Windows operating systems offer different cryptographic security technologies for data such as Encrypted File System and BitLocker. The capabilities of these tools depend on the presence of hardware modules for example, the Trusted Platform Module. The aim of this project is to complete a comprehensive study of the usage made of these technologies so that step-by-step tutorials can be prepared to guide users of these technological tools as per requirements.

Statistical Characterization of Real-World Distributed System Workloads

Enquisite Software provides analytical services to corporate clients through the “mining” of data detailing how the clients’ customers interact with the clients’ Internet points-of-presence. To facilitate this work, Enquisite collects a continuous stream of client interaction data, which is stored in an Enquisite-held repository, over which Enquisite clients perform on-demand queries for business actionable knowledge.