Call for Participation


2nd ACM/IEEE International Conference on
Distributed Smart Cameras (ICDSC-08)

September 7-11, 2008
Stanford University, California, USA

Panel Discussion

The pressing question facing many developers of systems and algorithms in vision has been where the new business opportunities are for cameras and distributed vision networks, and when meaningful applications based on vision will be introduced in the living and work environments of people. There appears to be no lack of technological tools today as many sensors and processors are commodities off-the-shelf, and there has been a plethora of concepts developed over the years about smart environments that sense and interact with their occupants. The novelties offered by multi-camera networks and the maturity of embedded processing technology can act as two driving forces that further motivate development of real-life and real-time applications. The panel aims to explore some key questions about the opportunities and challenges in developing such vision-based applications, including issues and intricacies related to the introduction of vision sensors in untested market environments. Another area of discussion is how "smart cameras" can play a role in defining new applications or enhancing the existing ones. Topics relating to cost-benefit tradeoffs in system design, scalability, and deployment will be discussed.

Tuesday, September 9

13:30 - 15:30 Hewlett 201

ICDSC-08 Panel
Business opportunities in multi-camera and embedded vision

Panel Members

  • David Forsyth [Univ. of Illinois, Urbana-Champaign]
  • Andrea Cavallaro [Queen Mary Univ. of London]
  • Steve Goldband [Stanford Center on Longevity]
  • Gianfranco Dorreto [GE Research]
  • Danny Petkevich [Texas Instruments]
  • Wei-Chao Chen [Nokia Research]
  • Mehran Farimani [Percipo]
  • Gary Bradksi [Willow Garage]


  • Tsuhan Chen [CMU]
  • Brad Stenger [Wired Magazine]
  • Hamid Aghajan [Stanford]

Aims and Topics

  • Opportunities for meaningful and real-life applications
  • Interplay of existing concepts and available technology
  • Challenges in vision processing: How smart cameras can play a role
  • How research can help address real needs of applications

    Vision to Multiple Smart Cameras

    • Can a paradigm change happen in computer vision based on distributed smart cameras?
    • What are the implications of:
      • Single camera -> Multiple cameras
      • Centralized -> Distributed processing
      • Specialized cameras -> webcams
      • Raw image transmitted, stored -> Images used, destroyed
      • Can find help for flaws / robustness issues of vision processing?


    • What are the impediments to adopting existing vision-based application concepts for practical implementation and use?
    • Where are the killer apps based on vision? Are there other real applications beyond security?

    Vision Processing

    • Is a key disabling factor for vision the issue of robustness of operation, such that it hinders the creation of any meaningful product or market?
    • How can multiple cameras help to take robustness to an acceptable level consistently? What would be the cost-benefit implications for such redundancy?

    Smart Cameras

    • What is the main motivation / justification for Smart cameras? Is it communication bandwidth saving? Scalability of the application? Pervasion of embedded processors?
    • How about for small-scale applications? Why smart cameras would help in home-based or office-based applications? Will it be an ease of deployment motivation?


    • Can a change of paradigm in security and surveillance (starting processing at the cameras) provide opportunities for novel design and additional features (scalability) in these applications?
    • What are the challenges in such transition? Are compromises to be made or can the new design surpass the legacy systems in all performance factors?

    Smart Environments

    • Any near-term applications? Will they be more likely a comfort enhancement application or a productivity enhancement application?
    • For smart environments applications, are privacy concerns (actual and perceived) a major impediment to adoption? How does the tradeoff between privacy concerns and value of the service play out for users in different applications?
    • Besides privacy, what are the other issues in using cameras in smart environments (intuitive use, adaptation to user preferences, unobtrusive and seamless design, etc.)? How can those be addressed by the research community?

    Role of Research

    • What are the most effective roles researchers in academia can play to help develop concepts and methods that are practical, or address issues challenging the market adoption of the technology?
    • Should research focus on developing methods for applications which can be made practical using available technology, or envision new concepts that may demand new technology? In other words, should technology drive concept or the other way around?