We are here to help you complete your graduate program and acquire the necessary skills for success after graduation.
Yasaman Sadat Sefidgar, MSc Thesis Defence, Computing Science
3:00 PM - 5:00 PM
March 11, 2014M.SC. THESIS DEFENCE Yasaman Sadat Sefidgar M.Sc. University of British Columbia, Canada, 2012 B.Sc. Sharif University of Technology, Iran, 2009 Tuesday, March 11th, 2014 3:00 p.m. TASC1 9204 East Title DISCRIMINATIVE KEY-SEGMENT MODEL FOR INTERACTION DETECTION Abstract Automatic activity detection in videos has several applications in human-computer interaction, visual surveillance, and video retrieval. The task, at its core, requires expressive models of activities. The models that represent activities as arrangements of key components are generally more descriptive and robust to challenges such as occlusion, clutter, and high intra-class variability. They can thus lead to improved classification performance. Following this idea, we model human-object interactions as sequences of locally discriminative temporal segments capturing objects’ appearance and their interrelations. In a two-stage pipeline, we coarsely localize humans and objects in long videos. We then more closely examine their content using our key-segment model trained in the latent SVM framework. We evaluate our approach on VIRAT Ground Dataset Release 2.0 for detecting instances of human-vehicle interactions. Results show that our key-segment model significantly outperforms the common global Bag of Words approach. M.Sc. Examining Committee: Dr. Greg Mori, Senior Supervisor Dr. Anoop Sarkar, Supervisor Dr. Ze-Nian Li, Examiner Dr. Brian Funt, Chair
Jared Fath, MSC Defence, Earth Sciences
2:00 PM - 5:00 PM
March 14, 2014No Description
Saeed Saberi Modaber, PhD Thesis Defence, Physics
11:00 AM - 2:00 PM
March 17, 2014DEPARTMENT OF PHYSICS THESIS DEFENCE DOCTOR OF PHILOSOPHY Monday, March 17, 2014 11:00 AM Room P8445.2, SFU Department of Physics Candidate: Mr. Saeed Saberi Modaber MSc, BSc, Sharif University of Technology Examining Committee Chair: Dr.Jeffrey McGuirk Senior Supervisor: Dr.Eldon Emberly Supervisor: Dr. John Bechhoefer Supervisor: Dr. Eric Cytrynbaum, University of British Columbia SFU Examiner: Dr. Erikur Palsson, Department of Biological Sciences External Examiner: Dr. Anirvan Sengupta, Rutgers University Models for Protein Localization in Bacteria and Chromatin Structure in Eukaryotes In the first part of this thesis, we focus on the problem of how proteins localize within the cytoplasm of bacteria. Experimentally, it is found that proteins that have attractive interactions are able to localize to one or both of the poles in cylindrical bacteria. We put forward a model that only relies on the aggregating tendency of proteins and the occlusion by the bacterial nucleoid. Monte-Carlo simulations enabled us to find the stable and metastable localization patterns patterns, allowing us to explore the phase space of parameters in our model. Our findings provide an explanation for the different patterns observed for PopZ localization in Escherichia coli and Caulobacter crescentus as well as misfolded proteins. We find the kinetics of expressing proteins has a crucial role: unipolar patterning is an energetically favorable state while other polar patterns can be achieved at higher rates of protein expression. Using sets of different GFP tagged aggregating proteins we are able to experimentally test this prediction and alter the localization from unipolar to bipolar simply by increasing the rate of expression. In the second part of the thesis, we consider the structuring of the DNA within the eukaryotic nucleous and its associated proteins. A new high-throughput experimental technique, Hi-C, is able to measure the looping frequencies between all parts of the genome. We transform the measured data and are able to extract a distance independent free energyby subtracting out the background free energy of interactions due to the polymer nature of DNA. Our mean-field model quantifies the interaction strengths between chromatin factors and loops along the chromosomes in a protein pairwise interaction matrix Jμ, υ. Since the Hi-C data carries different biases, using our approach we are able to assess the best sets of corrections that lead to the free energy having the most mutual information with the underlying chromatin profiles. Further to this, we use Principal Component Analysis (PCA) to identify the frequent modes of genome wide looping. Hence we are able to correlate these with known domain structures such as boundaries between active and silent regions of the genome.
Career Advising in Residence: Support for Spouses, Partners and Students March 07, 2014
Career Services is offering free career advising in Residence this term. Spouses and partners of grad students are welcome to book appointments.
Dean of Graduate Studies Awards for Excellence March 07, 2014
Nominate your fellow SFU faculty and staff members for the Dean of Graduate Studies Awards for Excellence.