aerial-sfu

Welcome to the Office of Graduate Studies & Postdoctoral Fellows

Future Students

Our internationally recognized graduate programs will expand your intellectual horizons, strengthen your skill set and enhance your professional network.

Current Students

We are here to help you complete your graduate program and acquire the necessary skills for success after graduation.

Student's researching

Upcoming Events

  • Nataliya Shapovalova, PhD Defence, Comp Sci
    10:00 AM - 12:00 PM
    November 27, 2014
    Ph.D. Thesis Defence   NATALIYA SHAPOVALOVA   B. Sc ., National Technical Univ "Kharkiv Polytechnic Institute", Ukraine, 2006 M . Sc .,  Heriot -Watt Univ, United Kingdom, 2009   Thursday November 27, 2014 10:00 a.m. TASC1 9204 West   TOWARDS ACTION RECOGNITION AND LOCALIZATION IN VIDEOS WITH WEAKLY SUPERVISED LEARNING   Human behavior understanding is a fundamental problem of computer vision. It is an important component of numerous applications, such as human-computer interaction, sports analysis, video search, and many others. In this thesis we work on the problem of action recognition and localization, which is a crucial part of human behavior understanding. Action recognition explains what a human is doing in the video, while action localization indicates where and when in the video the action is happening.   We focus on two important aspects of the problem: (i) capturing intra-class variation of action categories and (ii) inference of action location. Manual annotation of videos with fine-grained action labels and spatio-temporal action locations is a nontrivial task, thus employing weakly supervised learning is of interest. Real-life actions are complex, and the same action can look different in different scenarios. A single template is not capable of capturing such data variability. Therefore, for each action category we automatically discover small clusters of visually similar examples. A separate classifier is learnt for each cluster, so that more class variability is captured. Furthermore, we establish a direct association between a novel test example and examples from training data and demonstrate how metadata can be transferred to test examples.   Weakly supervised learning for action recognition and localization is another challenging task. It requires automatic inference of action location for all the training videos during learning. Initially, we simplify this problem and try to find discriminative regions in videos that lead to better recognition performance. The regions are inferred in a manner such that they are visually similar across all the videos of the same category. Ideally, the regions should correspond to the action location; however, there is a gap between inferred discriminative regions and semantically meaningful regions representing action location. To fill the gap, we incorporate eye gaze data to drive the learning. This allows inferring regions that are both discriminative and interpretable. Furthermore, we use the inferred regions and learnt action model to assist top-down eye gaze prediction.         Ph.D. Examining Committee: Dr. Greg Mori, Senior Supervisor Dr. Anoop Sarkar, Supervisor Dr. Mark Drew, Internal Examiner Dr. Jim Little, External Examiner Dr. Brian Funt, Chair  
  • Abdullah Aldhamin, MSc Thesis Defence, Computing Science
    10:00 AM - 12:00 PM
    November 27, 2014
    M.SC. THESIS DEFENCE Abdullah Aldhamin B .Sc. Computer Science (Honours), King Fahd University of Petroleum and Minerals, Saudi Arabia, 2007 Thursday, November 27 th , 2014 10:00 a.m. SURREY CAMPUS Room 3260 Title FLASH STORAGE MANAGEMENT ALGORITHM FOR LARGE-SCALE HYBRID STORAGE SYSTEMS Abstract Flash-based solid-state drives (SSDs) have led to significant innovations in storage systems architecture. However, due to their special design and architecture characteristics, they are not considered as cost-effective and immediate replacement of traditional hard disk drives (HDDs) for large-scale storage systems. Thus, how can we best utilize this technology to build an efficient hybrid storage system remains a research challenge. We propose a real-time dynamic programming algorithm, called Flash Storage Management (FSM) algorithm, to address this challenge. We implement the proposed FSM algorithm in an event-driven simulator. Our evaluation results indicate that the proposed algorithm outperforms other algorithms proposed in the literature, especially for the read-intensive workloads. In addition, the proposed FSM algorithm achieves better data swapping rate compared to other algorithms proposed in the literature, which indicates that the FSM algorithm works better in maintaining the lifetime of the SSD devices. M.Sc. Examining Committee: Dr. Mohamed Hefeeda, Senior Supervisor Dr. Raed AlShaikh, Supervisor Dr. Jiangchuan Liu, Examiner Dr. Joseph Peters, Chair
  • Jeffrey Ovens PhD Defence Chemistry
    2:00 PM - 6:00 PM
    November 27, 2014
    No Description
  • Download .ics