!!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!! !!!
-------------------------------------------------------------------------------
                                       
                                       
                   B O S T O N   U N I V E R S I T Y
                                       
                      Computer Science Department
                                       
                          C O L L O Q U I U M

          Recognition, Resolution, and Complexity of Objects 
                 Subject to Affine Transformations

                             Margrit Betke 
			   
		      Computer Science Department
			    Boston College


   			  Wednesday,  April 5 
			       3:00 PM 
                       (Coffee served at 2:45PM)
                  [Please note the non-standard time]

                                       
                         Seminar Room / MCS 135
                                       


    The problem of recognizing objects in images is examined from a
    physical perspective using the theory of statistical estimation.
    Focusing first on objects that occlude zero-mean scenes, we derive two
    useful descriptors from the objects' Fisher information that are
    independent of noise level.  The first is a generalized coherence
    scale that has great practical value because it corresponds to the
    width of the object's autocorrelation peak under affine transformation
    and so provides a physical measure of the extent to which an object
    can be resolved under affine parameterization. The second is a scalar
    measure of an object's complexity, which is invariant under affine
    transformation and can be used to quantitatively describe the
    ambiguity level of a general 6-dimensional affine recognition problem.
    Specifically, this measure of complexity has a strong inverse
    relationship to the level of recognition ambiguity.  We then develop a
    statistically optimal method for recognizing objects in scenes with
    zero-mean background, which is ``information-conserving,'' because it
    uses all the measured data pertinent to the object's recognition and
    exploits the continuous variations in shading that characterize the
    object but that are neglected in edge-based recognition methods.  We
    adapt our method to address the more general case of an object
    occluding a non-zero mean scene and apply it to recognize objects
    imaged in thousands of complex real-world scenes.  We show that the
    level of recognition ambiguity decreases exponentially with increasing
    object and scene complexity.  Ambiguity is then avoided by
    conditioning the permissible range of template complexity above
    apriori thresholds.


    Host:  Stan Sclaroff (sclaroff@cs.bu.edu)

-------------------------------------------------------------------------------
For colloquium info, including directions, see http://cs-www.bu.edu/colloquium
-------------------------------------------------------------------------------