The Conference on Computer and Robot Vision
Montréal, Quebec, May 7-9, 2014
Welcome to the home page for CRV 2014, the Eleventh Conference on Computer and Robot Vision. CRV is an annual conference which is held jointly with the Graphics Interface (GI) and Artificial Intelligence (AI) conferences. A single registration covers attendence in all three conferences. Please see the AI/GI/CRV general conference website for more information.
The CRV proceedings are published through the IEEE Computer Society's Conference Publishing Services (CPS). Accepted papers to CRV will be submitted to appear on the IEEE Xplore Digital Library.
- March 17th, 2014: The joint conference Registration Page is now active. Please register before the early-bird deadline of April 7th, 2014, if you plan to attend CRV.
- March 17th, 2014: The CRV Program is online.
- February 1st, 2014: A special extension until February 13th, 2014 for robotics papers only is announced (details on the Submissions tab).
- January 28th, 2014: The submission deadline has been extended until February 4th, 2014.
- December 11th, 2013: The submission site is now open and accepting contributions.
- November 18th, 2013: The CRV 2014 Program Committee has been finalized. See the Organizers tab.
- October 21st, 2013: Our two Keynote Speakers are confirmed to be Raffaelo D'Andrea from ETH Zurich and Frank Dellaert from Georgia Tech!
- October 14th, 2013: Announcing 4 special Research Symposia sessions at CRV this year. We have invited 16 professors from across Canada to participate as speakers. See the Symposia tab for more information.
- September 3rd, 2013: The CRV 2014 website is live!
|Paper submission (extended)||February 4 2014|
|Special robotics extension||February 13 2014|
|Acceptance/rejection notification||February 28 2014|
|Revised camera-ready papers||March 7 2014|
|Early registration||April 7 2014|
|Joint Conference||7-9 May 2014|
In 2004, the 17th Vision Interface conference was renamed the 1st Conference on Computer and Robot Vision. Therefore, CRV currently represents 27 years of outstanding academic content hosted within Canada. CRV is sponsored by the Canadian Image Processing and Pattern Recognition Society (CIPPRS).
CRV 2014 Program
Day 1: Wednesday, May 7th, 2014, U-de-M Andre-Aisenstadt Building
8:30-9:00 Joint Conference Welcome Session
9:00-10:30 Oral Paper Session - Navigation I
- Optimizing Camera Perspective for Stereo Visual Odometry.
Valentin Peretroukhin, Jonathan Kelly and Timothy D. Barfoot.
- Towards Estimating Bias in Stereo Visual Odometry.
Sara Farboud-Sheshdeh, Timothy D. Barfoot and Raymond H. Kwong.
- Stix-Fusion: A probabilistic Stixel integration technique.
Maximilian Muffert, Nicolai Schneider and Uwe Franke.
- Towards Full Omnidirectional Depth Sensing using Active Vision for Small Unmanned Aerial Vehicles.
Adam Harmat and Inna Sharf.
- Autonomous Maritime Landings for Low-cost VTOL Aerial Vehicles.
Kevin Ling, Derek Chow, Arun Das, and Steven L. Waslander.
10:30-11:00 Coffee Break
11:00-12:00 Keynote Speaker 1
- Raffaello D'Andrea, ETH Zurich
Sensing and Closing the Loop
12:00-14:00 Lunch (on your own)
14:00-15:30 Invited Symposium: Assistive Technologies
- Joelle Pineau, McGill University
Learning Socially Adaptive Navigation Strategies for Assistive Robots : Lessons from the SmartWheeler Project
- Ian Mitchell, University of British Columbia
Development of Shared Control Policies for Safe Wheelchair Navigation of Older Adults with Cognitive and Mobility Impairments
- Francois Michaud, Université de Sherbrooke
Interactive Robotics and Rehabilitation
15:30-16:00 Coffee Break
16:00-16:30 Invited Symposium: Assistive Technologies (continued)
- Jesse Hoey, University of Waterloo
General-Purpose Cognitive Assistive Technologies: Progress and Challenges
16:30-17:30 Oral Paper Session: Navigation II
- Self-Localization at Street Intersections.
Giovanni Fusco, Huiying Shen, Vidya Murali and James M. Coughlan.
- Adaptive Robotic Contour Following from Low Accuracy RGB-D Surface Profiling and Visual Servoing.
Danial Nakhaeinia, Pierre Payeur and Robert Laganière.
- Speed Daemon: Experience-based Mobile Robot Speed Scheduling.
Chris J. Ostafew, Jack Collier, Angela P. Schoellig and Timothy D. Barfoot.
- Image Retrieval using Landmark Indexing for Indoor Navigation.
Dwaipayan Sinha, Mirza Tahir Ahmed and Michael Greenspan.
Day 2: Thursday, May 8th, 2014, U-de-M Andre-Aisenstadt Building
9:00-10:15 Keynote Speaker 2
- Frank Dellaert, Georgia Tech.
Factor Graphs for Fast and Scalable 3D Reconstruction and Mapping
10:15-10:30 Oral Paper Session: Image Processing
- Generalized Exposure Fusion Weights Estimation.
Mohammed Elamine Moumene, Rachid Nourine and Djemel Ziou.
10:30-11:00 Coffee Break
11:00-12:00 Invited Symposium - Shape and Segmentation I
- James Elder, York University
Generative Models of Shape
- Kaleem Siddiqi, McGill University
12:00-12:30 Oral Paper Session: Shape
- 3D Scan Registration Using Curvelet Features.
Siddhant Ahuja and Steven L. Waslander.
- Building better formlet codes for planar shape.
Alex Yakubovich and James H. Elder.
12:30-14:00 Lunch (on your own)
14:00-15:00 Invited Symposium: Activity Recognition and Surveillance I
- Greg Mori, Simon Fraser University
Discriminative Latent Variable Models for Human Action Recognition
- Richard Wildes, York University
Spatiotemporal features for the detection of anomalous behaviour, actions and tracking
15:00-15:30 Oral Paper Session: Activity Recognition
- Sign Language Fingerspelling Classification from Depth and Color Images using a Deep Belief Network.
Lucas Rioux-Maldague and Philippe Giguerre.
- Multi-Task Learning of Facial Landmarks and Expression.
Terrance DeVries, Kumar Biswaranjan and Graham W. Taylor.
15:30-16:00 Coffee Break
16:00-17:00 Invited Symposium: Activity Recognition and Surveillance II
- Guillaume-Alexandre Bilodeau, École Polytechnique de Montréal
Improving visual surveillance using infrared and visible imagery
- Pierre-Marc Jodoin, Université de Sherbrooke
Video surveillance and UAV’s: challenges and future works
17:00-17:30 Oral Paper Session: Visual Attention
- Exploring Underwater Environments with Curiosity.
Yogesh Girdhar and Gregory Dudek.
- Visual Saliency improves Autonomous Visual Search.
Amir Rasouli and John K. Tsotsos.
Day 3: Friday, May 9th, 2014, Polytechnique Lassonde Building
9:00-10:00 Invited Symposium: Shape and Segmentation II
- Olga Veksler, Western University
Simple Priors for Image Segmentation
- Yuri Boykov, Western University
Combinatorial optimization for higher-order segmentation functionals: Entropy, Color Consistency, Curvature, etc.
10:00-10:30 Oral Paper Presentations: Segmentation
- Segmenting Objects in Weakly Labeled Videos.
Mrigank Rochan, Shafin Rahman, Neil D.B. Bruce and Yang Wang.
- Grid Seams: A fast superpixel algorithm for real-time applications.
Parthipan Siva and Alexander Wong.
10:30-11:00 Coffee Break
11:00-12:30 Invited Symposium - Field Robotics
- Gregory Dudek, McGill University
Sensor-based systems for terrestrial field robotics
- Steven Wasslander, Waterloo University
Advanced small rotorcraft autonomy
- Ralf Bachmayer, Memorial University of Newfoundland
Navigation, sensing and control of unmanned marine vehicles in an ocean environment
- John Tsotsos, York University
Tractability and Attention: Key Roles in Robotic Visual Search
12:30-14:00 Lunch (provided by conference at Poly Student Caf.)
14:00-15:30 Poster Session at Lassonde 3rd floor atrium (candy and chocolate at 3:30)Please see listing of posters below.
16:00-17:30 Oral Paper Session: Applications
- Direct Matrix Factorization and Alignment Refinement: Application to Defect Detection.
Zhen Qin, Peter van Beek and Xu Chen.
- Identification of morphologically similar seeds using multi-kernel learning.
Xin Yi, Mark Eramian, Ruojing Wang and Eric Neufeld.
- The Range Beacon Placement Problem for Robot Navigation.
River Allen, Neil MacMillan, Dimitri Marinakis, Rahnuma Islam Nishat, Rayhan Rahman, and Sue Whitesides.
- Trajectory Inference using a Motion Sensing Network.
Doug Cox, Darren Fairall, Neil MacMillan, D. Marinakis, David Meger, Saamaan Pourtavakoli, and Kyle Weston.
- Decentralized Cooperative Localization for Heterogeneous Multi-Robot System Using Split Covariance Intersection Filter.
Thumeera R. Wanasinghe and George K. I. Mann and Raymond G. Gosine.
- Asymmetric rendezvous at sea.
Malika Meghjani, Florian Shkurti, Juan Camilo Gamboa Higuera, Arnold Kalmbach, David Whitney, G. Dudek.
- A More Robust Feature Correspondence for More Accurate Image Recognition.
Shady Y. El-Mashad and Amin Shoukry.
- Projected Barzilai-Borwein Method with Infeasible Iterates for Nonnegative LeastSquares
K. Fraser, D. V. Arnold and G. Dellaire.
- Camera Matrix Calibration Using Circular Control Points and Separate Correction
of the Geometric Distortion Field.
Victoria Rudakova and Pascal Monasse.
- Computer Vision-Based Identification of Individual Turtles using Characteristic
Patterns of Their Plastrons.
Trevor Beugeling and Alexandra Branzan-Albu.
- “Drums”: a Middleware-Aware Distributed Robot Monitoring System.
Valiallah (Mani) Monajjemi, Jens Wawerla and Richard Vaughan.
- Scale-space and Wavelet Decomposition based Face Recognition using Nearest
Farhana Hoque and Liang Chen.
- Using Gradient Orientation to Improve Least Squares Line Fitting.
Tomislav Petkovic and Sven Loncaric.
- Trajectory Estimation Using Relative Distances Extracted from Inter-Image
Marten Wadenbäck and Anders Heyden.
- A Proof-Of-Concept Demonstration of Visual Teach and Repeat on a
Quadrocopter Using an Altitude Sensor and a Monocular Camera.
Andreas Pfrunder, Angela P. Schoellig and Timothy D. Barfoot.
- Outdoor ice accretion estimation of wind turbine blades using computer vision.
M. Akhloufi, N. Benmesbah.
- A Meta-Technique for Increasing Density of Local Stereo Methods through Iterative Interpolation and Warping.
Aniket Murarka and Nils Einecke.
- MDS-Based Multi-Axial Dimensionality Reduction Model For Human Action Recognition.
Redha Touati and Max Mignotte.
- Indoor Scene Recognition With a Visual Attention-Driven Spatial Pooling
Tarek Elguebaly and Nizar Bouguila.
- Automated Door Detection with a 3D-Sensor.
Sebastian Meyer zu Borgsen, Matthias Schopfer, Leon Ziegler and Sven Wachsmuth.
- Photon Detection and Color Perception at Low Light Levels.
Mehdi Rezagholizadeh and James J. Clark.
- Toward a Unified Framework for EMG Signals Processing and Controlling an
Guillaume Durandau and Wael Suleiman.
- Metadata-weighted Score Fusion for Multimedia Event Detection.
Scott McCloskey and Jingchen Liu.
- Robust detection of paint defects in moulded plastic parts.
Cole Tarry, Micheal Stachowsky and Medhat Moussa.
- Multiple Feature Fusion in the Dempster-Shafer Framework for Multi-Object
Dorra Riahi and Guillaume-Alexandre Bilodeau.
- Vision-Based Qualitative Path-Following Control of Quadrotor Aerial Vehicle with
Speed-Up Robust Features.
Trung Nguyen, George K. I. Mann and Raymond G. Gosine.
- Construction of a mean surface for the variability study of the cornea.
Arnaud Polette, Edouard Auvinet, Jean-Luc Mari, Isabelle Brunette, and Jean Meunier.
- A New Fitness Based Adaptive Parameter Particle Swarm Optimizer.
Sohail Akhtar, Eihab Abdel-Rahman and Abdul-Rahim Ahmad.
- An integrated bud detection and localization system for application in greenhouse
Cole Tarry, Patrick Wspanialy, Matthew Veres and Medhat Moussa.
- N-gram Based Image Representation And Classification Using Perceptual Shape
Albina Mukanova, Qigang Gao and Gang Hu.
- Interactive teleoperation interface for semi-autonomous control of robot arms.
Camilo Perez Quintero, Romeo Tatsambon Fomena, Azad Shademan, Oscar Ramirez and Martin Jagersand.
- Trinocular Spherical Stereo Vision for Indoor Surveillance.
Michel Findeisen and Gangolf Hirtz.
- Autonomous lecture recording with a PTZ camera while complying with
Dries Hulens, Toon Goedemé, and Tom Rumes.
- 3D reconstruction by fusioning shadow and silhouette information.
Rafik Gouiaa and Jean Meunier.
- President: Gregory Dudek, McGill University
- Treasurer: John Barron, Western University
CRV 2014 Co-Chairs
- David Meger, McGill University
- Michael Langer, McGill University
CRV 2014 Program Committee
- Allili, Mohand Said, Université du Québec en Outaoauis, Canada
- Allison, Robert, York University, Canada
- Andreopoulos, Alexander, IBM Research, Canada
- Bachmayer, Ralf, Memorial University of Newfoundland, Canada
- Barfoot, Tim, University of Toronto, Canada
- Barron, John, University of Western Ontario, Canada
- Bergevin, Robert, Université Laval, Canada
- Bilodeau, Guillaume-Alexandre, École Polytechnique Montréal, Canada
- Boulanger, Pierre, University of Alberta, Canada
- Boyd, Jeffrey E., University of Calgary, Canada
- Carneiro, Gustavo, University of Adelaide, Australia
- Clark, James, McGill University, Canada
- Clausi, David, University of Waterloo, Canada
- Cobzas, Dana, University of Alberta, Canada
- Collier, Jack, DRDC Suffield, Canada
- Derpanis, Kosta, Ryerson University, Canada
- Dudek, Gregory, McGill University, Canada
- Elder, James, York University, Canada
- Eramian, Mark, University of Saskatchewan, Canada
- Ferworn, Alexander, Ryerson University, Canada
- Fieguth, Paul, Waterloo, Canada
- Forstner, Wolfgang, Berlin University, Germany
- Funt, Brian, Simon Fraser University, Canada
- Giguère, Philippe, Laval University, Canada
- Greenspan, Michael, Queens University, Canada
- Gupta, Kamal, Simon Fraser University, Canada
- Hoey, Jesse, University of waterloo, Canada
- Hogue, Andrew, University of Ontario Institute of Technology, Canada
- Hoover, Randy C., South Dakota School of Mines and Technology, USA
- Ivekovic, Spela, University of Strathclyde, United Kingdom
- Jagersand, Martin, University of Alberta, Canada
- Jenkin, Michael, York University, Canada
- Jiang, Hao, Boston College, USA
- Jodoin, Pierre-Marc, Université de Sherbrooke, Canada
- Laganière, Robert, University of Ottawa, Canada
- Lalonde, Jean-Francois, Laval University, Canada
- Lang, Jochen, SITE, University of Ottawa, Canada
- Langer, Michael, McGill University, Canada
- Laurendeau, Denis, Laval University, Canada
- Li, Howard, University of New Brunswick, Canada
- Little, Jim, University of British Columbia, Canada
- Malik, Shahzad, University of Toronto, Canada
- McCloskey, Scott, Honeywell Labs, USA
- Meger, David, McGill University, Canada
- Meunier, Jean, Universite de Montreal, Canada
- Michaud, Francois, Université de Sherbrooke, Canada
- Mignotte, Max, Universite de Montreal, Canada
- Miller, Gregor, University of British Columbia, Canada
- Mori, Greg, Simon Fraser University, Canada
- Pal, Christopher, École Polytechnique Montréal, Canada
- Pradalier, Cédric, Georgia Tech. Lorraine, France
- Rekleitis, Yiannis, McGill University, Canada
- Sattar, Junaed, University of British Columbia, Canada
- Scharfenberger, Christian, University of Waterloo, Canada
- Siddiqi, Kaleem, McGill University, Canada
- Smith, Steven, University of Waterloo, Canada
- Sohn, Gunho, York University, Canada
- Spetsakis, Minas, York University, Canada
- Stilla, Uwe, Technische Universitaet Muenchen, Germany
- Tong, Chi Hay, University of Oxford, United Kingdom
- Tsotsos, John, York University, Canada
- Vaughan, Richard, Simon Fraser University, Canada
- Veksler, Olga, University of Western Ontario, Canada
- Wang, Yang, University of Manitoba, Canada
- Waslander, Steven, Waterloo University, Canada
- Wong, Alexander, Waterloo University, Canada
- Xiao, Yijun, University of Edinburgh United Kingdom
- Yang, Herb, University of Alberta, Canada
- Yilmaz, Alper, Ohio State University, USA
- Young, James Everett, University of Manitoba, Canada
- Zelek, John, University of Waterloo Ontario, Canada
- Zhang, Hong, University of Alberta, Canada
CRV 2014 will feature two exciting keynote talks.
Wednesday, May 7th, 2014, 11am
Raffaello D'Andrea, ETH Zurich
Title: "Sensing and Closing the Loop"Abstract: In this talk I will discuss how sensing was used in the various machines and systems I've been involved with over the years, in research, education, competitions, business, and the arts.
Bio: Spanning academics, business and the arts, Raffaello D'Andrea's career is built on his ability to bridge theory and practice: He is Professor of Dynamic Systems and Control at the Swiss Federal Institute of Technology (ETH) in Zurich, where his research redefines what autonomous systems are capable of. He is also co-founder of Kiva Systems (recently acquired by Amazon), a robotics and logistics company that develops and deploys intelligent automated warehouse systems. In addition, he is an internationally-exhibited new media artist, best known for the Robotic Chair (Ars Electronica, ARCO, London Art Fair, National Gallery of Canada) and Flight Assembled Architecture (FRAC Centre, France).
Thursday, May 8th, 2014, 9am
Frank Dellaert, Georgia Tech. University
Title: Factor Graphs for Fast and Scalable 3D Reconstruction and MappingAbstract: Simultaneous Localization and Mapping (SLAM) and Structure from Motion (SFM) are important and closely related problems in robotics and vision. I will show how both SLAM and SFM instances can be posed in terms of a graphical model, a factor graph, and that inference in these graphs can be understood as variable elimination. The overarching theme of the talk will be to emphasize the advantages and intuition that come with seeing these problems in terms of graphical models. For example, while the graphical model perspective is completely general, linearizing the non-linear factors and assuming Gaussian noise yields the familiar direct linear solvers such as Cholesky and QR factorization. Based on these insights, we have developed both batch and incremental algorithms defined on graphs in the SLAM/SFM domain. In addition to direct methods, we recently worked on efficient iterative methods that use subgraphs of these factor graphs as pre-conditioners in a conjugate gradient scheme. Finally, we are now looking into how optimal control can be seamlessly integrated with the estimation algorithms for use in autonomous vehicles.
Bio: Frank Dellaert is a Professor in the School of Interactive Computing at the Georgia Institute of Technology. He joined Georgia Tech in 2001 after obtaining a Ph.D. from Carnegie Mellon's School of Computer Science, where he worked with Hans Moravec, Chuck Thorpe, Sebastian Thrun, and Steve Seitz. Before that, he also obtained degrees from Case Western Reserve University and The Catholic University of Leuven, in Belgium. His research is in the areas of Robotics and Computer vision. He is particularly interested in graphical model techniques to solve large-scale problems in mapping and 3D reconstruction. You can find out about his research and publications at http://www.cc.gatech.edu/~dellaert. The GTSAM toolbox which embodies many of the ideas his group has worked on in the past few years is available for download at http://tinyurl.com/gtsam
CRV 2014 will feature four exciting symposia on a selection of subtopics relatied to computer and robotic vision. Titles and abstracts and links to speaker web pages are given below. Please see the Program for the detailed schedule.
Assistive Technologies/HealthWed. May 7, 14:00-15:30 and 16:00-16:30
Chair: Joelle Pineau
"Learning Socially Adaptive Navigation Strategies for Assistive Robots : Lessons from the SmartWheeler Project"
Abstract: A key skill for mobile robots is the ability to navigate efficiently through their environment. In the case of social or assistive robots, this involves navigating through human crowds. Typical performance criteria, such as reaching the goal using the shortest path, are not appropriate in such environments, where it is more important for the robot to move in a socially acceptable manner. In this talk I will describe new methods based on imitation and reinforcement learning which we have developed to allow robots to achieve socially adaptive path planning in human environments. Performance of these methods will be illustrated using a smart power wheelchair developed in our group, called the SmartWheeler.
University of British Columbia
“Development of Shared Control Policies for Safe Wheelchair Navigation of Older Adults with Cognitive and Mobility Impairments”
Abstract: Advances in sensor and robotics technology may soon translate into practical smart wheelchairs for the growing population of users with a combination of mobility and cognitive or sensory impairments. CanWheel is an interdisciplinary, cross-Canada project investigating how power wheelchairs are used now, how they could be used better, and how they could be better. This talk will concentrate on the engineering side of CanWheel, including earlier research which informed our focus on the cognitively impaired population; recent work in which we seek to understand how members of this population might best collaborate with a robotic wheelchair to accomplish their goals and to publicly document a robot's view of the driving behavior of this population and the environment in which they live; and challenges to working in this assistive technology domain.
Université de Sherbrooke
“Interactive Robotics and Rehabilitation”
Abstract: Designing an interactive and mobile robot is a challenging area of robotics. It involves dealing with motion, interaction modalities (vision, audition, touch) and interfaces, allowing robots to interact in natural ways with people and the environment. Such capabilities can be useful in the field of rehabilitation. During this talk, we introduce IRL-1, a humanoid platform equipped with an expressive face, an orientable head, a 8-microphone array, two compliant arms with four degrees of freedom (DOF) each and grippers, all installed on an omnidirectional, non-holonomic and compliant mobile platform. We then illustrate how IRL-1’s capabilities are used in the development of interactive technologies in rehabilitation, such as an omnidirectional powered wheelchair, force-guided platform, appearance-based SLAM, and an augmented multimodal (vision, audition and touch) teleoperation interface.-->
University of Waterloo
“General-Purpose Cognitive Assistive Technologies: Progress and Challenges”
Abstract: Over the past decade, our multidisciplinary team have been building intelligent systems to provide assistance for persons with dementia wishing to age-in-place. These systems sense the environment with cameras and other sensors, and use probabilistic and decision-theoretic controllers to compute prompting and assistance strategies. In this talk, I will give an overview of our work, showing examples from multiple tasks and trials that we have performed. I will discuss our efforts to generalise these cognitive assistive technologies so they can be be set up and managed by ``informal caregivers'' in a DIY ("do-it-yourself") fashion. I will then discuss our recent work on endowing these assistive technologies with the ability to reason about the personal (socio-cultural and affective) identities of users, and about the affective delivery of interventions. -->
Activity Recognition (I) and Surveillance (II)Thurs. May 8, (I) 14:00-15:00 and (II) 16:00-17:00
Chair: Guillaume-Alexandre Bilodeau
Greg Mori, Simon Fraser University
“Discriminative Latent Variable Models for Human Action Recognition”
Abstract: Developing algorithms to interpret scenes of human activity involves a number of related tasks including human detection, tracking, and action recognition. These tasks are intertwined, information from one can provide assist in solving others. In this talk we will describe discriminative latent variable models to address these tasks together, focusing on the latent SVM / max-margin hidden conditional random field. These methods can be used for jointly recognizing actions and spatio-temporally localizing them in videos. Models for human-human and human-object interactions will be presented. We will present methods for group activity recognition, with holistic analysis of entire scenes of people interacting and taking different social roles.
Richard Wildes, York University
"Spatiotemporal features for the detection of anomalous behaviour, actions and tracking"
Abstract: A spatiotemporal feature set will be presented that is applicable to a wide range of video surveillance tasks. The feature set is derived by representing visual spacetime with a distributed oriented energy representation. This representation systematically exposes the structure of visual spacetime in terms of local, multiscale orientation and thereby supports integrated analysis of spatial appearance and dynamic image properties. The utility of the approach will be illustrated in application to detection of anomalous behaviour, human actions and tracking.
Université de Sherbrooke
“Video surveillance and UAV’s: challenges and future works”
Abstract: So far, UAVs and drones have been mainly used for military applications. But things are about to change drastically with new market-ready low-cost UAVs. It is expected that UAVs will soon become ubiquitous and will produce the same kind of revolution the IP cameras did 10-15 years ago. In this presentation, I will review some current and future surveillance applications related to low-altitude UAVs. I will demonstrate how well-known computer vision problems such as image registration, optical character recognition, change detection, scene understanding, shape-from-motion, camera calibration, and pedestrian detection can be deeply redefined when used in the context of a low-altitude UAV. This presentation is focused on unsolved issues, UAV-specific problems and future works for the computer vision community.
École Polytechnique de Montréal
“Improving visual surveillance using infrared and visible imagery”
Abstract: Multimodal visual surveillance has drawn a lot of attention recently because using other modalities in addition to the standard visible spectrum imagery allows improving significantly object detection, recognition, and activity recognition tasks. In this talk, we will focus on the combination of visible and infrared video data for improving people segmentation. We will first explain the benefit of combining visible and infrared images and discuss the related research challenges. Then, we will go through the steps required to obtain improved human segmentations in videos, namely, initial object segmentation using background subtraction, tracking to gather temporal information on the people in the scene, registration of each person silhouette, which in the end will lead to a final refined segmentation of each person. We will conclude by discussing some remaining challenges to study.
Shape (I) and Segmentation (II)Thurs. May 8 (I) 11:00-12:00, and Fri. May 9 (II) 9:00-10:00
Chair: James Elder
James Elder, York University
“Generative Models of Shape“
Abstract: While shape plays an important role in leading recognition frameworks, it is normally used in a discriminative way. Fully generative models of shape have potential advantage in supporting recognition while also facilitating segmentation in cluttered scenes. I will review a selection of generative models but will focus on an approach originating with the GRID framework of Grenander et al (2007), used to model incremental growth in medical imagery, which has the significant advantage of preserving topology. I will show how this framework can be used to model natural shapes as the result of sequence of localized deformations called formlets, applied to an embryonic elliptical form. Although growth is strictly sequential in nature, we show how relaxing this restriction results in a natural hierarchical representation that corresponds closely with our intuitive notion of parts and introduces opportunities for parallelizing and thus speeding shape coding and decoding.
Kaleem Siddiqi, McGill University
Abstract: I shall provide an overview of our work in modeling close to parallel curves that turn slowly, with applications to hair modeling and pattern analysis in computer vision and medical imaging. Our strategy has been to attach a frame field to the relevant data and to then characterize its local curvature. This provides fruitful ways of parametrizing and quantifying the local “shape" of these complex patterns.
Olga Veksler, Western University
“Simple Priors for Image Segmentation”
Abstract: In this talk we will describe several simple priors for image segmentation: compact, star, tiered, and convex. We call these priors `simple' as they are not specific to a particular object, such as 'a guitar' or 'a bicycle'. Rather they encourage the object to obey some simple generic geometric shape, such as a trapezoid, or discourage highly non-convex objects. Such geometrically simple objects are abundant in real-world or medical images. All of these priors can be incorporated into the commonly used energy-based segmentation framework. In particular, a typical energy function includes the appearance term and the length regularizer,in addition to the shape prior term. With the exception of the convex shape prior, all of these energies with the shape term included can be efficiently and globally optimized, either with a graph cut or dynamic programming. Optimization for the convex shape prior has to be approximated. Finally, we demonstrate the usefulness of shape priors for segmentation on real images.
Yuri Boykov, Western University
“Combinatorial optimization for higher-order segmentation functionals: Entropy, Color Consistency, Curvature, etc.”
Abstract: This talk discusses recent advances in image segmentation in the context of higher-order appearance and smoothness functionals. The most basic segmentation energies combine unary terms enforcing certain color appearance models and pair-wise terms enforcing boundary smoothness. If color models are known, the corresponding binary optimization problem can be globally minimized. Estimation of color models leads to NP-hard mixed optimization problems that are typically solved with iterative block-coordinate descent (Zhu-Yuille, Chan-Vese, GrabCut, etc.) sensitive to initialization. This talk motivates higher-order appearance functionals (e.g. entropy and color-consistency) that do not require explicit estimation of segment appearance models. We show that in many cases such energies can be minimized globally. For example, our approach allows replacing iterative “grabcut” technique with a one cut method finding a global minimum. We also discuss a general Trust Region approach for approximate minimization of other high-order appearance terms. Time permitting, we will also motivate higher-order boundary smoothness terms (e.g. curvature) and describe the corresponding state-of-the-art combinatorial optimization techniques.
Joint Symposium with the NSERC Canadian Field Robotics NetworkFri. May 9, 11:00-12:30
Chair: Gregory Dudek
Gregory Dudek, McGill University, NCFRN Scientific Lead and representing theme: LAND
"Sensor-based systems for terrestrial field robotics"
Abstract: In this talk I will briefly provide an overview and objectives of the NSERC Canadian Field Robotics Network, and then proceed to discuss a sampling of results from the sub-theme dealing with terrestrial robotics. This includes the use of image-based representations for navigation and localization, the use of unsupervised terrain classifiers based on haptic feedback for robot gait selection, cooperative camera targeting, and high dimensional path planning for complex robots. In these each of these domains we combine aspects of sensing, modeling, and environmental interaction targeted to challenging unconstrained environments.
Steven Waslander, University of Waterloo, theme leader: AIR
"Advanced small rotorcraft autonomy"
Abstract: Working on the Asctech Pelican platform throughout the NCFRN network has led to rapid development of autonomous capabilities for quadrotor vehicles. With contributions from McGill, SFU and Waterloo, we are seeking to demonstrate combined multi-camera visual state estimation and feature based mapping, vehicle control for autonomous docking and human-robot interaction algorithms at this year's field trials. The collaborative project builds on recent independent work by each group, and both past results and the current objectives will be described in this talk.
Ralf Bachmayer, Memorial University of Newfoundland, theme leader: WATER
"Navigation, sensing and control of unmanned marine vehicles in an ocean environment"
Abstract: As part of the NCFRN network, the water theme sub-group is working with several unmanned marine platforms, at the surface and under water. The AQUA platform, in various configurations, is suitable for applications in terrestrial, amphibious including surf-zone, and underwater environments. The SLOCUM underwater gliders are long endurance deep diving vehicles for oceanographic sensing and detection. The surface platforms consist of two catamaran type unmanned surface craft, a Clearpath Kingfisher vehicle and an AOSL-SeaCat, with a third vehicle type, a semisubmersible system under construction. This talk will introduce the platforms and give an overview of the various research topics addressed within the water theme, such as advances in the locomotion gaits for the aqua platform, sparse terrain aided navigation for underwater gliders, iceberg sensing an mapping concepts. An overview of the coordinated and cooperative control of surface, underwater and potentially aerial platforms for the purpose of the full data acquisition of and around an iceberg will be presented.
York University, representing theme: HUMAN
“Tractability and Attention: Key Roles in Robotic Visual Search”
Abstract: Visual Search for objects, locations or events of interest is a central capability for a robot with real-world utility. But this can not limited to yes-no detection; it must include an ability to measure, describe, and compare within the context of a task. We have been investigating this problem since the late 1980's and regardless of the prevailing trends in machine vision, learning, or robotics, have not found reason to ignore the roles of attention and of a deep understanding of the computational nature of the problem. This presentation will briefly trace our journey. Along the way, we emphasize a number of major points, including the roots of our approach in issues of tractability, the design and evaluation of our subsumptive search algorithm, the development of the AIM saliency model, the confounding nature of sensor bias, the integration of saliency within the object search algorithm, and the need for an overarching framework for attentive behavior, which we have named Cognitive Programs.
Links to Previous Conferences
This page archives historical content from past CRV meetings. A second source for some of this information is maintained at the CIPPRS website.
- CRV 2013: Regina, Saskatchewan, May 28-31, 2013
- CRV 2012: Toronto, Ontario, May-28-30, 2012
- CRV 2011: St-John's, Newfoundland and Labrador, May-25-27, 2011
- CRV 2010: Ottawa, Ontario, 31 May-2 June 2010
- CRV 2009: Kelowna, British Columbia, 25-27 May 2009
- CRV 2008: Windsor, Ontario, 28-30 May 2008
- CRV 2007: Montreal, Quebec, 28-30 May 2007
- CRV 2006: Quebec City, Quebec, 7-9 May 2006
- CRV 2005: Victoria, British Columbia, 9-11 May 2005
- CRV 2004: London, Ontario, 17-19 May 2004.
Description : Downtown Montéal from parc du mont Royal
Credit : © Tourisme Montréal