Skip to content
Department of Computer Science

Computer Science



  • 2017 Humanoid Application Challenge Winners

    The Autonomous Agents lab has just won first place in the 2017 Humanoid Application Challenge at this year's IEEE International Conference on Robotics and Systems (IROS), IEEE's flagship robotics conference. The Humanoid Application Challenge is intended to be more open-ended than most other robotics competitions, in that entries are judged on dimensions of effectiveness and innovation in a given theme rather than stating a precisely defined goal such as winning a soccer competition. This perspective encourages creative entries that cross boundaries and bring together work from many areas of artificial intelligence that are important to intelligent humanoid robots, including vision, speech understanding, coordination, reasoning, machine learning, and human-robot interaction.

    Click here for more information
  • Best Paper Award

    Congratulations to Mohammad Moein Almasi, MSc student with Dr. Hadi Hemmati, for winning the Best Paper Award in the 39th International Conference in Software Engineering (ICSE 2017)! 

    Click here for more information
  • New Program Approval Process

    We are moving away from the paper program approval forms you may have used in previous years. Instead, we will be using a fillable PDF form, available online under the "resources" link forms section, that you should fill in and e-mail to If you wish to have a face to face meeting please indicate this in your e-mail otherwise your form will simply be dealt with electronically and you will receive an e-mail once this has been done. 

    Click here for more information


  • M.Sc. Thesis Defense: Saliency Ranking using Deep Learning

    When: May 24, 2018 @ 1:00pm
    Where: E2-461 EITC

    Speaker: Mahmoud Kalash


    Salient object detection is a problem that has been considered in detail and many solutions proposed. In this thesis, we argue that work to date has addressed a problem that is relatively ill-posed. Specifically, there is not universal agreement about what constitutes a salient object when multiple observers are queried.

    This implies that some objects are more likely to be judged salient than others, and implies a relative rank exists on salient objects. The solution presented in this thesis solves this more general problem that considers relative rank, and we propose data and metrics suitable to measuring success in a relative object saliency landscape.

    First, a novel deep learning solution is proposed based on a hierarchical representation of relative saliency and stage-wise refinement. We also show that the problem of salient object subitizing can be addressed with the same network and our approach exceeds performance of any prior work across all metrics considered (both traditional and newly proposed).

    Furthermore, we present data, analysis and benchmark baseline results towards addressing the problem of salient object ranking. Methods for deriving suitable ranked salient object instances are presented, along with metrics suitable to measuring algorithm performance. In addition, we show how a derived dataset can be successively refined to provide cleaned results that correlate well with pristine ground truth. We also demonstrate the value of different rejection thresholds in determining exemplars suitable for training and evaluation, demonstrating the superior performance of a large dataset that may have some minor labeling noise over smaller extant datasets.

    Finally, we provide a comparison among prevailing algorithms that address salient object ranking or detection to establish initial baselines.

Social Media

© 2011 University of Manitoba Department of Computer Science
Back to top