For more Human Computer Interaction theses, click here

  • Fast 3D Reconstruction of Human Figures in Motion by Tashia Nasir.  Supervised By: Chris Joslin

Abstract

3D Reconstruction of Human Figures

Our goal was to develop a complete scheme for the dynamic volumetric 3D reconstruction of human figures. We aimed at achieving three key objectives that were accuracy, real-time acquisition of partial view 3D data followed by fast processing of data to produce a complete 3D model, and cost-effectiveness of the process. The research problems we encountered were the selection of appropriate camera technology and capturing environment to carry out the reconstruction. We designed a setup that consisted of multiple depth cameras placed around the object of interest at a fixed distance ranging between 0.75m to 1m. We placed the setup in a uniformly lit room and removed any glossy or reflective surfaces from the scene. We designed the system and setup based on the experimental findings we conducted to assess the conditions that aided the accomplishment of our set goals. The multiple depth cameras despite projecting IR did not interfere with each other due to the unique patterns they projected. An essential step of the reconstruction process was the alignment stage where the multiple scans of the target object were aligned for the final reconstruction. A research problem was the accuracy analysis of the various alignment algorithms when they were applied to the target scans individually or in combination. The effect of the initial orientation and increasing percentage of overlap between the consecutive scans on the alignment accuracy was an essential factor to analyze. The combinations of alignment algorithms that worked the best in all overlapping and initial orientation conditions were key point matching coarse alignment algorithms combined with the fine alignment algorithms. The increase in percentage overlap had a positive effect on the alignment while scan orientations other than being parallel to each other had a negative impact on the accuracy of the alignment. We presented a complete methodology where the reconstruction part itself took an average of three minutes conforming to our research goals of fast reconstructions some additional pre and post-processing steps added to the overall processing time for the reconstruction.We achieved the minimum error percentage of 4.67% in our final reconstructed 3D model under conditions where we had 50% overlap between individual scans and initial parallel orientation of the scans to be aligned.

  • Bendy: An Exploration into Gaming with Mobile Flexible Devices   by Jessica Lo, M.A.Sc. Supervised by: Audrey Girouard

Abstract 

We explore gaming with mobile flexible devices through a comprehensive study. We conducted two experiments with two commonly sized mobile devices. The first experiment created user-defined bend gestures from a set of tasks derived from gaming. The results suggest pairing opposing events by gesture location, little differences in gestures performed on two sized devices and interaction paradigms relating to the Simon Effect. The results of the first experiment were summarized and provided guidelines for the design of the game controls used in the second experiment. We implemented the user-defined gestures into 6 games using an interactive flexible prototype. Usability and user experience were evaluated and we find an overall preference for the small size, identified an important usability issue related to hand position and interaction paradigms regarding the Simon Effect and bimanual input. Finally, we propose a set of design recommendations for flexible device interactions and game designers.

  • “A Day in the Life of the Jos”: The Design of an Educational Game on Privacy by Christine Mekhail, M.A. 

    Supervised by: Sonia Chiasson

        

    ADITL Calendar showing active days with scenarios in red

     Abstract

    The need to educate children on protecting their online privacy and security is pressing, as more children are going online unsupervised. This thesis examines how to make privacy and security education relevant to children’s everyday life. We have partnered with MediaSmarts to create a new interactive educational game. As a starting point, we conducted a user study of an existing game, then we designed, prototyped, and evaluated a new game with children. In our new “A Day in the Life of the Jos” (ADITL) game, players follow Jo and Josie, a brother and sister, through their day and help them make positive privacy and security-related choices. Players
    then witness the outcomes of their decisions on the Jos’ lives as they unfold within the game. Evaluation results show that participants found ADITL appealing and that adding gamification elements encouraged engagement with the educational material and promoted interaction with the lesson

  • Creating 3D Human Character Mesh Prototypes from a Single Front-View Sketch by Shaikah Bakerman, M.C.S. Supervised by: Chris Joslin

Sketch by Geoffrey Datema ©

        Abstract

3D character modeling, a vital part in film and video game production, is a process that starts by blocking out the basic geometry of a character, which is then transformed into a detailed and enhanced mesh. This process can be a long and daunting task to novice modelers. As a result, extensive research in the area of sketch-based modeling has focused on finding solutions that facilitate this process.

We developed a sketch-based modeling system that constructs the basic 3D geometry of a human character mesh based on a single front-view sketch and minimal user interaction through a simple interface. The main objective of this system is to help novice modelers by automating the initial phase of the modeling process with the aid of an intuitive user interface, and to construct a basic mesh with suitable structure that conforms to common modeling techniques.

  • Investigating the Potential of Tabletop Natural User Interfaces Tools in Improving the Nunaliit Cybercartographic Atlas Framework by Omar Ibrahim Mahmoud Bani Taha , M.A. Supervised by: Taylor Fraser

 Abstract

This study presents the results of an on-going work, which offers a comprehensive qualitative analysis of natural hybrid interfaces (touch-based, gesture and tangible interfaces), as a human-computer interaction technique that has the potential to promote a methodological approach to the design of a new form of collaborative tabletop interface within the Nunaliit framework and its cybercartographic atlases.

The study provides empirical evidence for the feasibility and value of incorporating collaborative interactive large displays in a mapping creation process. The results of this study are based on a usability study comprised of twenty participants and semi-structured interviews with ten professionals from various fields and experiences. The study confirms the potential benefits and applicability of employing this novel approach as an alternative to the more conventional user interfaces that are currently in use. The study offers several insights and design guidelines which will be indispensable when implementing this novel interface, particularly in the new collaborative Nunaliit framework, and offer a new opportunity for cartographic mapping contexts in general.

  • Adding Cartoon-Like Motion to Realistic Animations by Rufino Ricardo Ansara, M.A.Sc. Supervised by : Chris Joslin

Walt Disney Animation Studios, Frozen. 2013.

       Abstract             

In comparison to traditional animation techniques, motion capture allows animators to obtain a large amount of realistic data in little time. The motivation behind our research is to try to fill the gap that separates realistic motion from cartoon animation. With this, classical animators can produce high quality animated movies (such as Frozen, Toy Story, etc.) and non-realistic video games in a significantly shorter amount of time. To add cartoon-like qualities to realistic animations, we suggest an algorithm that changes the animation curves of motion capture data by modifying local minima and maxima. We also propose a curve-based interface that allows users to quickly edit and visualize the changes applied to the animation. Finally, we present the results of two user studies that evaluate both the overall user satisfaction with the system’s functionality, interactivity and learning curve, and the animation quality.