Advances in Robot Vision - From Domestic Environments to Medical Applications
IROS 2004 workshop
WWF4 Advances in Robot Vision -
From Domestic Environments to Medical Applications

 

Registration deadline: August 20 through IROS 2004 web

Date:
Wednesday, September 29: 09:45-16:45 (full-day)
Place:
Meeting Room 4

Organizers:
Danica Kragic and Henrik I Christensen
Centre for Autonomous Systems
Royal Institute of Technology
Sweden

Computer vision is gaining significant importance as a cheap, passive, and information-rich sensor in research areas such as unmanned vehicles, medical robotics, human-machine interaction, autonomous navigation, robotic manipulation and grasping. One of the current trends is to build complex computer vision systems based on the integration of simple and tractable methods. However, most of the systems we know today are designed for special purposes such as object manipulation, medical or underwater applications, etc. Some of the natural question one may pose are: i) What are the basic building blocks of such systems, ii) What is the overlap between different systems and is there a set of cues/methods commonly used accross different research disciplines? iii) Are there some similarities regarding the underlying integration and control strategies?

This workshop aims at presenting state-of-the-art computer vision research and related problems such as algorithm design and modeling, underlying control methodologies, visual cues and their integration, real-time aspects, vision for autonomus navigation and cognitive aspects.

It is our hope that this workshop will bring together computer vision scientists from different disciplines and open for a vivid discussion on future milestones and wide use of computer/robot vision systems.

Presenters (Schedule):

  • Jana Kosecka, (George Mason University, USA)

    Geometric and Appearance Based Methods for Visual Model Acquisition and Localization


  • Francois Chaumette and Eric Marchand, (IRISA/INRIA, Rennes, France)

    Features Extraction and Tracking for Visual Servoing Purpose
    marchand

  • Ezio Malis and Selim Benhimane (INRIA, Sophia Antipolis, France)

    Unified Framework for Real-time Visual Tracking and Servoing
    malis

  • Koichi Hashimoto, (Tohoku University, Japan)

    A Hierarchical Control Architecture for High-Speed Visual Servoing
    koichi

  • Gregory D. Hager, (The Johns Hopkins University, MD, USA)

    Navigating Inner Space: 3-D Assistance for Minimally Invasive Surgery

  • Ameesh Makadia and Kostas Daniilidis, (University of Pennsylvania, USA)

    Robot Localization without Correspondence


  • Darius Burschka , (The Johns Hopkins University, MD, USA)

    Vision-Based SLAM with Standard Off-Shelf Cameras
    darius

  • Nick Barnes, Gareth Loy and Luke Fletcher, (National ICT / The Australian National University, Australia)

    Robot Vision for Driver Support Systems
    loy

  • Markus Vincze, Michael Zillich and Wolfgang Ponweiser (Technical University, Wien, Austria)

    A Software Framework to Integrate Vision and Reasoning Components for Cognitive Vision Systems
    markus

  • Danica Kragic , Marten Bjorkman and Henrik I Christensen (Centre for Autonomous Systems (CAS), KTH, Sweden)

    Issues and Strategies for Robotic Object Manipulation in Domestic Settings
    dani


    Motivation and objectives:

    Until recently the use of vision in real-time applications has required simplification in the used methods to adhere to complexity limitations. The advanced in available computer power, and the introduction of new methods in terms of sensors, control and recognition has allowed design of a new generation of vision systems. The operative term here is systems. Today, vision offers a flexible modality for generation of feedback to robot control systems. Earlier research has had to rely on relatively simple methods visual cues or other sensory modalities for generation of systems. The recent availability of more flexible sensory modalities has generated an entire class of flexible systems and opened up a number of new applications.

    The objective of this workshop is to bring together researchers from a wide variety of application domains to discuss:

  • Vision for control
  • Vision for recognition
  • Vision for interpretation

    Through a careful analysis of the involved methods, it is of interest to identify basic techniques in terms of

  • Sensor design
  • Cue generation
  • Cue integration
  • Recognition
  • Control generation and servoing

    Each of the application domains offer different constraints and impose specific requirements, yet many of the basic methods are the same. To make progress, it is of interest to identify the underlying methods. In many cases the methods may already be available in the computer vision community, but they are not in general use in the robotics community. There is consequently a need to bring the communities of robot and computer vision to a joint appreciation of the value of systems; where there is a need to consider all aspects from perception to action generation. The objective of the workshop is to provide an overview of recent robotic applications of vision and identify the underlying methods. Based on the set of basic method it is of interest to discuss future challenges in terms of systems oriented vision.

    For the authors:
    **** Sample files ****** Sample Files


  • Danica Kragic's homepage
  • Henrik Christensen's homepage