\magnification=1200
\baselineskip=20pt
\nopagenumbers
\font\big=cmr12 scaled \magstep2

\centerline{\bf STANFORD UNIVERSITY}
\centerline{\bf DEPARTMENT OF STATISTICS}
\centerline{\big DEPARTMENTAL SEMINAR}
\bigskip

\baselineskip=12pt
\centerline{4:15 p.m., Tuesday, November 26, 2002}
\centerline{Sequoia Hall Room 200}
\centerline{(Cookies at 3:45 in 1st Floor Lounge)}

\bigskip
\baselineskip=15pt
\centerline{\sl Assaf Zeevi}
\centerline{\sl Columbia University}
\bigskip
\centerline{\bf The Hough Transform Estimator}
\bigskip

 The Hough transform is the mainstream computer vision algorithm used
 to  detect and fit lines to a set of planar points, e.g.,
a noisy image.  In this talk we pursue
 a ``statistical view'' of the Hough transform. We study asymptotic
 properties  of the corresponding Hough transform estimator,
establishing consistency,
 rates of convergence and characterizing  the limiting distribution.
 We will also examine robustness-related properties of this class
of estimators.  Several numerical examples serve to illustrate the
various properties of
 the Hough transform  estimator.   Regarding the problem of testing for
multiple lines,
 we establish some  preliminary asymptotic results that  point out some
 theoretical as well as  practical limitations of the methodology.

(Joint work with Alexander Goldenshluger, Haifa University)

\bye