EURASIP Journal on Applied Signal Processing

  Special Issue on

  Tracking in Video Sequences of Crowded Scenes

  Call for Papers

 Object tracking in live video is an enabling technology that is in strong
 demand by large application sectors, such as video surveillance for security
 and behavior analysis, traffic monitoring, sports analysis for enhanced TV
 broadcasting and coaching, and human body tracking for human-computer
 interaction and movie special effects.

 Many techniques and systems have been developed and demonstrated
 for tracking objects in video sequences. The specific goal of this
 special issue is to provide a status report regarding the state
 of the art in object tracking in crowded scenes based on the video
 stream(s) of one or more cameras. The objects can be people, animals,
 cars, and so forth. The cameras can be fixed or moving. Moving cameras
 may pan, tilt, and zoom in ways that may or may not be communicated
 to the tracking system.

 All papers submitted must address at least the following two issues:

  o Processing of live video feeds

    For many applications in surveillance/security and TV sports broadcasting,
    the results of processing have value only if they can be provided to the
    end user within an application-defined delay. The submitted papers should
    present algorithms that are plausibly applicable to such incremental
    (^Ócausal^Ô) processing of live video feeds, given suitable hardware.

  o Handling of crowded scenes

    Crowded-scene situations range from relatively simple (e.g., players
    on a planar field in a soccer match) to very difficult (e.g., crowds
    on stairs in an airport or a train station). The central difficulties
    in crowded scenes arise from the constantly changing occlusions of any
    number of objects by any number of other objects.

 Occlusions can be resolved to some degree using a single video stream.
 However, many situations of occlusion are more readily resolved by the
 simultaneous use of several cameras separated by wide baselines. In
 addition to resolving ambiguities, multiple cameras also ease the
 exploitation of 3D structure, which can be important for trajectory
 estimation or event detection.

 Topics of interest include principles and evaluation of relevant end-to-end
 systems or important components thereof, including (but not limited to):

  o Handling of occlusions in the image plane in single-camera scenarios
  o Handling of occlusions in a world coordinate system (3D, possibly
    degenerated to 2D) in single- or multicamera scenarios
  o Fusion of information from multiple cameras and construction of integrated
    spatiotemporal models of dynamic scenes
  o 3D trajectory estimation
  o Tracking of multiple rigid, articulated, or nonrigid objects
  o Automatic recovery of camera pose from track data
  o Detection and recognition of events involving multiple objects
   (e.g., offside in soccer)

 Authors should follow the EURASIP JASP manuscript format
 described at the journal site http://alert.hindawi.com/journals/asp/
 Prospective authors should submit an electronic copy of their
 complete manuscript through the EURASIP JASP's manuscript
 tracking system at journal's web site, according to the
 following timetable.

  Manuscript Due           October 1, 2005
  Acceptance Notification  February 1, 2006
  Final Manuscript Due     May 1, 2006
  Publication Date         3nd Quarter, 2006

GUEST EDITORS:

 Jacques G. Verly, Department of Electrical Engineering and Computer Science,
 University of Liège (ULg), Sart Tilman, Building B28, 4000 Liège, Belgium;
 jacques.verly@ulg.ac.be

 John MacCormick, Microsoft Research, Silicon Valley, 1065 La Avenida Mountain
 View, CA 94043, USA; jmacc@microsoft.com

 Stephen McKenna, Division of Applied Computing, University of Dundee, Dundee
 DD1 4HN, Scotland, UK; stephen@computing.dundee.ac.uk

 Justus H. Piater, Department of Electrical Engineering and Computer Science,
 University of Liège (ULg), Sart Tilman, Building B28, 4000 Liège, Belgium;
 justus.piater@ulg.ac.be